Method for analyzing influence of railway tunnel construction on urban underground water

Abstract

The invention discloses a method for analyzing the influence of railway tunnel construction on urban underground water, which belongs to the field of railway tunnel construction and comprises the following steps: S1, acquiring and preprocessing multi-source data; s2, constructing a geology-hydrology-construction-remote sensing multi-dimensional feature vector, and extracting an underground water disturbance index; s3, establishing a stress-seepage-heat conduction three-field coupling model; s4, analyzing a three-field coupling effect based on the three-field coupling model, and outputting a comprehensive risk level of the influence of tunnel construction on underground water; s5, fusing a comprehensive risk level by utilizing a GIS spatial analysis function, and combining real-time data to realize spatial visualization and graded early warning of underground water influence; and S6, geological parameters of the three-field coupling model are updated in real time, and an optimal construction parameter combination is searched. By adopting the method for analyzing the influence of the railway tunnel construction on the urban underground water, the scientificity and engineering practicability of analyzing the influence of the railway tunnel construction on the urban underground water are remarkably improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of railway tunnel construction, and particularly relates to a method for analyzing the influence of railway tunnel construction on urban groundwater. BACKGROUND

[0002] With the acceleration of urbanization, railway tunnel engineering is increasingly deep into the city dense area, and the influence of construction disturbance on urban groundwater system has become a key problem of engineering safety and environmental protection. The traditional analysis method has the following technical bottlenecks:

[0003] 1. Insufficient precision of multi-source data fusion: The existing technology mainly relies on a single data source (such as borehole survey or InSAR (synthetic aperture radar interferometry) monitoring, which is difficult to fully reflect the multi-dimensional influence of tunnel construction on groundwater. Among them, the traditional borehole sampling density is low (usually one borehole is arranged every 50-100m), which is difficult to capture the spatial variation of rock mass damage and permeability coefficient. The error of permeability coefficient estimation based on standard penetration test (SPT) can be up to ±30%, and it cannot reflect the permeability mutation caused by blasting damage. And the existing InSAR displacement solving method (such as PS-InSAR) is disturbed by atmospheric delay, thermal noise and other interference, which leads to large error in extracting construction disturbance signal. For example, in a subway tunnel engineering, the traditional filtering method mistakenly identifies atmospheric disturbance as construction settlement, causing risk misjudgment.

[0004] 2. Insufficient intelligence of face state recognition: The weathering degree (WD) and groundwater state (GC) of the tunnel face are the key factors affecting the seepage field. For this, early reliance on manual geological recording is time-consuming and subject to subjective experience. Later, a structure surface recognition method based on point cloud clustering or region growing was developed, but it needs to manually set parameters, and the robustness is poor. For example, the misjudgment rate of the clustering algorithm based on DBSCAN is more than 15% in the dense crack area.

[0005] 3. Incomplete multi-physical field coupling modeling: The existing model does not fully consider the multi-physical field interaction effect of construction disturbance, such as stress-seepage coupling simplification: the elastoplastic model based on Mohr-Coulomb criterion ignores the influence of damage on permeability. In a certain railway tunnel engineering, the permeability attenuation in the blasting damage area (which actually reduces to 1 / 5 of the original rock) is not considered, resulting in a permeation flow prediction error of more than 50%. Temperature effect is missing: the change of pore water pressure caused by tunnel construction heat source (such as mechanical friction heat) is not included in the analysis. Theoretical research shows that the increase of 5℃ in temperature can increase the seepage flow by 12% to 18%, but the existing model generally ignores this effect. Dynamic damage evolution is ignored: the spatio-temporal evolution process of damage factor (D) is not coupled with the seepage field in real time, and the dynamic change of permeability caused by support lag cannot be reflected.

[0006] 4、Existing risk assessment systems have the following defects: (1) Static index system: risk classification based on a single parameter (such as seepage flow threshold) cannot reflect the coupling effect of multiple factors. In a certain tunnel project, although the seepage flow did not exceed the threshold, stress concentration led to sudden gushing, exposing the limitations of traditional methods. (2) Weak spatial analysis capability: the integration of GIS (Geographic Information System) and BIM (Building Information Modeling) only stays at static layer superposition, lacking dynamic risk field simulation. For example, in a certain subway project, the risk superposition effect in the fault intersection area was not timely identified, leading to a lag in emergency plans. (3) Insufficient real-time performance: the monitoring data and model updating cycle are long (usually in days), which cannot capture the instantaneous risk changes during construction. In a certain tunnel, the seepage flow increased by 3 times within 1 hour after blasting, but the existing system failed to provide timely warning.

[0007] 5、Traditional construction parameter adjustment relies on experience and lacks a closed loop of "monitoring-analysis-optimization", with the following defects: (1) Lack of parameter sensitivity analysis: the optimization of support pressure and excavation rate lacks quantitative basis. In a certain tunnel project, blindly increasing support pressure led to a 20% increase in cost, but the risk reduction was less than 10%. (2) Poor dynamic adaptability: changes in geological conditions during construction (such as sudden faults) cannot trigger automatic parameter adjustment. In a certain railway tunnel, the increase in overbreak due to the failure to timely adjust the blasting parameters intensified the disturbance of groundwater. (3) Insufficient data reliability: monitoring data tampering risks (such as sensor failure and false alarms) affect optimization decisions.

[0008] In summary, the core defects of existing technologies are insufficient multi-source data fusion, incomplete physical process modeling, static risk assessment, and experiential construction optimization. These problems lead to insufficient analysis precision of the impact of railway tunnel construction on urban groundwater, making it difficult to meet the dual demands of engineering safety and environmental protection. SUMMARY

[0009] The purpose of the present application is to provide a method for analyzing the impact of railway tunnel construction on urban groundwater, which intelligently integrates multi-modal data, enhances features through deep learning, models thermal-hydraulic-mechanical three-field coupling, integrates GIS-BIM-blockchain early warning, and optimizes digital twins, systematically solving the above technical bottlenecks and significantly improving the scientificity and engineering practicability of the analysis method.

[0010] To achieve the above purpose, the present application provides a method for analyzing the impact of railway tunnel construction on urban groundwater, comprising the following steps:

[0011] S1, multi-source data acquisition and preprocessing, constructing a multi-dimensional spatio-temporal data set containing surface fissures, geology and hydrology, and construction disturbance;

[0012] S2, by multi-dimensional feature fusion algorithm integration of pre-processed multi-source data, construct geology-hydrology-construction-remote sensing multi-dimensional feature vector, and utilize the improved RNPC-net model to identify the state of the working face, and extract the underground water disturbance index;

[0013] S3, based on the elastic-plastic damage theory and seepage mechanics, a stress-seepage-thermal conduction three-field coupling model is established, and the boundary conditions of the stress-seepage-thermal conduction three-field coupling model are set based on the underground water disturbance index;

[0014] S4, based on the three-field coupling model, the three-field coupling effect is analyzed, and based on the comprehensive risk assessment model, the comprehensive risk grade of the influence of tunnel construction on underground water is output;

[0015] S5, the comprehensive risk grade is fused by using the GIS spatial analysis function, and the spatial visualization and grading early warning of the influence of underground water are realized in combination with real-time data;

[0016] S6, the geological parameters of the three-field coupling model are updated in real time by using the extended Kalman filtering algorithm, the construction process virtual mapping is constructed by combining the digital twin technology, and the optimal construction parameter combination is searched, so as to form a monitoring-analysis-optimization closed loop control.

[0017] Therefore, the railway tunnel construction influence on urban underground water analysis method has the beneficial effects that:

[0018] 1, multi-source data acquisition and preprocessing: solve the problem of traditional single data source, improve the data fusion accuracy, can capture the spatial variation of rock mass damage and permeability coefficient, and reduce the error of construction disturbance signal extraction;

[0019] 2, the multi-dimensional feature fusion algorithm and the improved RNPC-net model are used to identify the state of the working face, the weathering grade of the working face and the probability distribution of the underground water state are output, the intelligentization and accuracy of the working face state identification are improved, and the time-consuming and subjective experience influence of manual recording are reduced;

[0020] 3, based on the elastic-plastic damage theory and seepage mechanics, a stress-seepage-thermal conduction three-field coupling model is established, considering the construction heat source temperature effect, the time and space evolution of the damage factor and the real-time coupling of the seepage field, solving the problem of traditional model multi-physical field coupling modeling not comprehensive, improving the prediction accuracy of seepage quantity;

[0021] 4, a comprehensive risk assessment model is constructed, a comprehensive risk grade is output based on multiple risk indexes, a risk-measure mapping relationship is established, and the problems of traditional static index system and weak spatial analysis capability are solved, which can reflect the coupling effect of multiple factors;

[0022] 5. The GIS spatial analysis function is used to fuse and integrate the risk level, combined with BIM and Unity3D platform to realize the visualization of groundwater influence space and hierarchical warning, and to improve the timeliness of risk response and dynamic risk field modeling capability;

[0023] 6. The extended Kalman filter algorithm is used to update the model geological parameters in real time, combined with digital twin technology to construct virtual mapping of the construction process and search for the optimal construction parameter combination, form a monitoring-analysis-optimization closed loop control, solve the problem of traditional construction parameter adjustment depending on experience, realize parameter dynamic optimization, improve construction efficiency and safety, and reduce cost.

[0024] The technical solutions of the present application will be further described in detail below with the aid of the accompanying drawings and examples. BRIEF DESCRIPTION OF DRAWINGS

[0025] Figure 1 A flowchart of a method for analyzing the influence of railway tunnel construction on urban groundwater. DETAILED DESCRIPTION

[0026] In order to make the purpose, technical solutions and advantages of the embodiments of the present application clearer, the embodiments of the present application will be further described in detail below with the aid of the accompanying drawings and examples. It should be understood that the specific embodiments described herein are only used to explain the embodiments of the present application and do not limit the embodiments of the present application. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor fall within the scope of the present application. The examples of the embodiments are shown in the accompanying drawings, wherein the same or similar reference numerals represent the same or similar elements or elements having the same or similar functions throughout.

[0027] It should be noted that the terms "comprise" and "have" and any variations thereof are intended to cover non-exclusive inclusion, for example, a process, method, system, product or server comprising a series of steps or units does not have to be limited to only those steps or units clearly listed, but can include other steps or units not clearly listed or inherent to the process, method, product or device.

[0028] The embodiments of the present application will be described in detail below with the aid of the accompanying drawings.

[0029] As shown in Figure 1 A method for analyzing the influence of railway tunnel construction on urban groundwater, comprising the following steps:

[0030] S1, multi-source data acquisition and preprocessing, constructing a multi-dimensional spatio-temporal data set containing surface fissures, geology and hydrology and construction disturbance;

[0031] Step S1 specifically comprises the following steps:

[0032] S11, collect multi-source data: obtain Sentinel-1A / B satellite C-band remote sensing data through EGMS to obtain surface fissure data; at the same time, obtain drilling data, geological survey report, underground water level monitoring data, permeability coefficient test results and pore water pressure distribution data to form geological hydrological data; and obtain construction process parameters and working face images to form construction disturbance data;

[0033] In step S11, the drilling data is obtained as follows: drill holes are arranged every 50-100 m along the tunnel axis, and the drilling depth penetrates 10-20 m below the tunnel floor, and the lithology distribution data is obtained by collecting core samples for lithology classification;

[0034] At the same time, the number of blows N SPT of 30 cm penetration is obtained by standard penetration test, and the effective internal friction angle of rock mass

[0035]

[0036] The construction process parameters include TBM tunneling parameters and blasting construction parameters, and the TBM tunneling parameters include cutterhead torque M t , advance speed v p and shield pressure P shield ; the blasting construction parameters include single-hole charge Q b , blast center distance r b , blast time interval and ground vibration speed v b .

[0037] S12, multi-source data preprocessing:

[0038] For remote sensing data: first, according to the orbital geometric parameters of Sentinel-1 satellite, the LOS displacement sequence of the ascending and descending data of the remote sensing data is projected to the vertical direction to eliminate the radar incidence angle θ deviation, and the effective monitoring points are retained by coherence threshold screening, and the corrected vertical displacement sequence d vert is output:

[0039]

[0040] In the formula, d LOS represents the original LOS displacement time sequence;

[0041] Then, the effective monitoring points are screened by coherence threshold, the principal component analysis-independent component analysis method decomposition technology is used to separate the construction disturbance component and the natural settlement component in the displacement signal, and the construction disturbance displacement sequence d excav is extracted;

[0042] For borehole data: particle size is tested by sieving and sedimentation method to obtain clay content, and then the permeability coefficient K = K0·10 -3 is preliminarily estimated based on the clay content, K0 represents the reference permeability coefficient related to the clay content, and finally the preliminary estimated permeability coefficient K is calibrated through the pressure water test to obtain the anisotropic permeability coefficient tensor K i of each layer, wherein the formula for calculating the permeability coefficient K is as follows:

[0043] At the same time, the fracture development degree is analyzed according to the borehole CT image, and the damage factor D is defined L fracture and L core respectively represent the surface crack length and the core length, and then the ordinary Kriging interpolation is used to generate the three-dimensional damage factor field D(x, y, z).

[0044] S2, integrating the preprocessed multi-source data through a multi-dimensional feature fusion algorithm, constructing a geological-hydrological-construction-remote sensing multi-dimensional feature vector, and using an improved RNPC-net model to identify the status of the working face and extract the underground water disturbance index;

[0045] Step S2 specifically includes the following steps:

[0046] S21, constructing a geological-hydrological-construction-remote sensing multi-dimensional feature vector:

[0047] S211, based on the corrected vertical displacement sequence d excav , the tunnel diameter R and the burial depth z0, calculating the settlement characteristic parameters, the settlement characteristic parameters including the volume loss rate V L and the slot width d;

[0048] The volume loss rate V L is calculated based on the Gaussian settlement model:

[0049]

[0050] In the formula, V s represents the volume of the settlement tank, and ω(x) represents the vertical displacement distribution function of the surface settlement tank, and ω max represents the maximum surface settlement, ω max =max(d excav ), l i represents the inflection point distance of the settlement tank, x represents the lateral coordinate; A t represents the cross-sectional area of the tunnel, and R represents the diameter of the tunnel;

[0051] The slot width d is calculated by the following formula:

[0052]

[0053] In the formula, i represents the inflection point distance; z0 represents the tunnel burial depth;

[0054] S212, constructing a feature vector: integrating geology and hydrology data, construction data, remote sensing data and surface fissure distribution data to obtain a feature matrix F=[K i ,n,D,v p ,M t ,V L ,d,ω max ]n represents the surface porosity;

[0055] S213, using the minimum-maximum standardization method to map each feature value in the feature matrix to the interval [0, 1] to obtain the normalized feature matrix F norm ;

[0056] S22, identifying the tunnel face: constructing an improved RNPC-net model and using the preprocessed tunnel face image, three-dimensional coordinates and feature matrix F norm to train the improved RNPC-net model until convergence, outputting the weathering grade probability distribution P(WD) and the groundwater state probability distribution P(GC):

[0057] The improved RNPC-net model adopts a three-branch mixed feature extraction module, which includes a two-dimensional image convolution branch with ResNet50 as the backbone, a three-dimensional point cloud processing branch and a construction parameter full connection branch, and fuses multi-source data input through an adaptive weight auxiliary classifier;

[0058] The cross-entropy loss function is set as follows:

[0059]

[0060] In the formula, : represents the loss value; y s represents the label vector; P s represents the distribution probability; S represents the total number of categories; s represents the category index;

[0061] S23, extracting groundwater disturbance indexes, wherein the groundwater disturbance indexes include a weathering comprehensive index WI and a groundwater activity intensity index GI;

[0062] Wherein, the weathering comprehensive index WI is calculated based on the weathering grade probability distribution P(WD):

[0063]

[0064] In the formula, w k represents the weight; P(WD k) represents the predicted probability of the kth weathering grade, k = 1, 2, 3, 4 respectively represent unweathered, slightly weathered, moderately weathered and strongly weathered;

[0065] And the groundwater state probability distribution P(GC) is mapped to a numerical index, and a groundwater activity intensity index GI is constructed:

[0066] GI = 4P(FW) + 3P(DW) + 2P(DP) + P(DR);

[0067] In the formula, P(FW), P(DW), P(DP) and P(DR) respectively represent the probability of water gushing at the working face, the probability of water dripping at the working face, the probability of water dripping at the working face, and the probability of water dripping at the working face.

[0068] S3, based on the elastoplastic damage theory and seepage mechanics, a stress-seepage-heat conduction three-field coupling model is established, and the boundary conditions of the stress-seepage-heat conduction three-field coupling model are set based on the groundwater disturbance index;

[0069] Step S3 specifically includes the following steps:

[0070] S31, elastoplastic damage stress field modeling:

[0071] S311, inversion of damage zone parameters, including blasting damage radius and anisotropic material parameter matrix [E d ,v d ], wherein the blasting damage radius is estimated by the following empirical formula, and is calibrated combined with the sound wave test result to determine the damage zone range:

[0072]

[0073] In the formula, R d represents the geometric parameter of the damage zone; k b represents a geological correction coefficient;

[0074] The damage zone elastic modulus E d in the anisotropic material parameter matrix [E d ,v d ] is calculated according to the following formula:

[0075] E d = E0(1-D);

[0076] In the formula, W0 represents the elastic modulus of the original rock;

[0077] The Poisson's ratio v d is calculated according to the following formula:

[0078] v d = v0+0.1D;

[0079] In the formula, v0 represents the Poisson's ratio of the original rock;

[0080] S312, the radial effective stress σ' of the elastic zone er and the tangential effective stress σ' eθ :

[0081]

[0082]

[0083] In the formula, P0 represents the original rock stress; r represents the inner wall of the tunnel; R i represents the inner radius of the tunnel;

[0084] Based on the Mohr-Coulomb criterion, the radial effective stress σ' of the plastic zone pr and the tangential effective stress σ' pθ are derived as follows:

[0085]

[0086] In the formula, represents the cohesion of the plastic zone, and c p represents the cohesion of the plastic zone; represents the soil pressure coefficient of the plastic zone, and R p represents the radius of the plastic zone; represents the internal friction angle of the plastic zone, and η represents the damage reduction coefficient;

[0087] S32, the seepage equation considering the temperature effect is constructed, and the boundary conditions are set, so as to obtain the pore water pressure distribution cloud map and the seepage flow Q, wherein the expression of the seepage equation is as follows:

[0088]

[0089] In the formula, v represents the seepage velocity vector; μ represents the dynamic viscosity of water; P w represents the pore water pressure; ρ represents the water density; g represents the gravitational acceleration; h represents the water head height; β represents the thermal expansion coefficient; T represents the temperature field distribution;

[0090] The boundary conditions are as follows: when the inner wall of the tunnel r=R i , the pore water pressure P w =P wi , P wi represents the water pressure in the tunnel; when the far field r=R w , P w =P w0 , R w represents the seepage influence radius, and P w0represents initial pore water pressure;

[0091] The seepage flow Q is calculated according to the following formula:

[0092]

[0093] In the formula, H represents the thickness of the aquifer.

[0094] S4, based on the three-field coupling model, analyze the three-field coupling effect, and based on the comprehensive risk assessment model, output the comprehensive risk grade of the influence of tunnel construction on groundwater;

[0095] Step S4 specifically includes the following steps:

[0096] S41, coupling the stress field and the seepage field to calculate the total stress:

[0097] r =σ' r +P w ,σ θ =σ' θ +P w ;

[0098] In the formula, σ r and σ θ respectively represent the radial total stress and the tangential total stress; σ' r and σ' θ respectively represent the stress field radial stress and the tangential stress, and σ' r =σ' er +σ' pr ,σ' θ =σ' eθ +σ' pθ ;

[0099] S42, introducing the construction heat source, calculating the temperature distribution through the heat conduction equation, and correcting the seepage parameters:

[0100]

[0101] In the formula, μ(T) represents the dynamic viscosity of water after temperature correction, and μ(T)=μ0exp(-k T (T-T0)), μ0 represents the dynamic viscosity of reference water, k T represents the temperature coefficient, and T and T0 respectively represent the real-time temperature and the initial temperature;

[0102] S43, setting the risk indicators as water gushing risk, settlement risk and extrusion risk, wherein the water gushing risk The settlement risk The extrusion risk Q th and d th are set threshold values, r​qd represents rock quality designation; a single risk factor vector [RI Q ,RI d , SPI] is obtained;

[0103] S44, a comprehensive risk index CRI is constructed:

[0104] CRI = ω Q RI Q + ω d RI d + ω SPI SPI;

[0105] In the formula, ω Q , ω d and ω SPI all represent weights, and ω Q + ω d + ω SPI = 1;

[0106] S45, the risk is divided into low risk, medium risk and high risk based on the comprehensive risk index CRI, and a risk-measure mapping relationship is established.

[0107] S5, the comprehensive risk level is fused by using the GIS spatial analysis function, and the spatial visualization and grading early warning of groundwater influence are realized in combination with real-time data;

[0108] Step S5 specifically includes the following steps:

[0109] S51, the geological parameters, the analysis results of the three-field coupling model and the real-time monitoring data are uniformly converted into ESRIShapefile format and projected into the UTM coordinate system;

[0110] S52, a thematic layer of the geological layer, the stress field and the seepage field is created in ArcGIS Pro;

[0111] S53, the inverse distance weighted method is adopted to interpolate the discrete risk points to generate a continuous risk surface:

[0112]

[0113] In the formula, R(x, y, z) represents the continuous risk value at the three-dimensional space coordinates (x, y, z); d i represents the spatial distance between the prediction point and the i th discrete risk monitoring point; CRI i represents the comprehensive risk index of the i th discrete risk monitoring point; I represents the total number of discrete risk monitoring points;

[0114] S54. Construct a BIM tunnel model based on tunnel geometry parameters, construction process parameters, and the analysis results of the three-field coupling model. Import the BIM tunnel model and comprehensive risk level data into the Unity 3D platform to achieve a dynamic correlation display between construction progress and risk evolution.

[0115] S55. Set graded warning rules: Level 1 warning conditions: Comprehensive risk index CRI ≥ 0.7 and surface settlement rate > 1mm / day, triggering a red warning and automatically starting the emergency grouting plan;

[0116] Level 2 warning conditions: 0.4≤CRI<0.7 or sedimentation rate>0.5mm / day, triggering a yellow warning and prompting increased monitoring frequency.

[0117] S6. Use the extended Kalman filter algorithm to update the geological parameters of the three-field coupling model in real time, combine it with digital twin technology to build a virtual mapping of the construction process, and search for the optimal combination of construction parameters to form a monitoring-analysis-optimization closed-loop control.

[0118] Step S6 specifically includes the following steps:

[0119] S61, using the extended Kalman filter to update the geological parameters of the three-field coupling model in real time, and output the updated damage factor D new and permeability coefficient K new ;

[0120] Define the state vector x = [D, K i ], observation vector represents the measured value of the vertical displacement of the ground surface, Q meas Representing the measured value of seepage flow, the following state equation and observation equation are established:

[0121] x k+1 =f(x k )+w k ;

[0122] z k =h(x k )+v k ;

[0123] Where x k+1 and x k denote the predicted value of the state vector at time k+1 and the estimated value of the state vector at time k respectively; f(·) denotes the state transfer function; w k represents state noise; z k represents the observation vector at time k; h(·) represents the observation function; v k represents the observation noise;

[0124] The state estimation value is iteratively calculated by using an extended Kalman filter algorithm:

[0125]

[0126] In the formula, and respectively represent the optimal estimation value of the state vector at time k and the state prediction value at time k based on the estimation at time k-1; K k represents the Kalman gain, and P k|k-1 represents the covariance matrix of the state prediction at time k based on the estimation at time k-1, represents the covariance matrix of the observation noise;

[0127] S62, the updated damage factor D new and the permeability coefficient K new are substituted into the three-field coupling model in step S3 to recalculate the stress field, the seepage field and the risk index;

[0128] S63, the calibrated results are compared with the measured data, the root mean square error is calculated to evaluate the precision of the three-field coupling model, and if the root mean square error is greater than 5%, the process returns to step S61 for secondary calibration;

[0129] S64, the construction parameters are optimized;

[0130] S641, a digital twin model is constructed: a tunnel construction digital twin is developed based on the Unity platform, a BIM tunnel model and real-time monitoring data are integrated, and 1:1 virtual mapping of the construction process is realized;

[0131] S642, a proximal strategy optimization algorithm is used to search for the optimal parameter combination and respectively represent the optimal support pressure and the optimal excavation rate.

[0132] Simulation experiment

[0133] Experimental conditions: software platform: COMSOL Multiphysics 6.0, coupled seepage field-stress field-temperature field module. Geological model: stratigraphy: divided into 3 layers, simulating typical urban unconsolidated sediments. Groundwater: initial water level depth 5m, porosity 0.25, permeability coefficient decreasing with depth. Tunnel: diameter 8m, buried depth 20m, excavated by TBM, support lagging behind the excavation face by 10m. Boundary conditions: the top is a free surface, and the bottom is a fixed water head boundary (water level depth 30m). Gradient pore water pressure is applied to the tunnel excavation face to simulate the change of permeability coefficient caused by construction disturbance (permeability coefficient in the disturbance zone increases by 1-3 times).

[0134] Table 1 stratum parameters

[0135]

[0136] Based on the above conditions, the influence of tunnel construction on urban groundwater is analyzed by using the method of the application and the traditional method (single Gaussian permeation model, ignoring stress-seepage coupling, assuming that the permeability coefficient is constant) respectively.

[0137] Table 2 seepage flow comparison

[0138]

[0139] It can be seen that the error of the predicted seepage flow of the application is 8.5%, and the error of the traditional method is 22.3%, which shows that the three-field coupling model of the application significantly improves the accuracy.

[0140] And the maximum water level drawdown predicted by the traditional method is 3.2m, and the influence radius is 50m, the maximum drawdown predicted by the application is 2.8m, and the influence radius is 42m, which is more consistent with the field monitoring (drawdown 2.9m, radius 40m), thereby proving the effectiveness of the application.

[0141] Finally, it should be noted that: the above examples are only used to illustrate the technical solutions of the application and not to limit them, although the application has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that: it can still modify or equivalently replace the technical solutions of the application, and these modifications or equivalent replacements cannot make the modified technical solutions deviate from the spirit and scope of the technical solutions of the application.

Claims

1. A method for analyzing the impact of railway tunnel construction on urban groundwater, characterized by: The following steps are involved: S1. Multi-source data collection and preprocessing to construct a multi-dimensional spatiotemporal dataset including surface fractures, geology, hydrology, and construction disturbances; S2. Integrate pre-processed multi-source data through a multi-dimensional feature fusion algorithm to construct a multi-dimensional feature vector of geology, hydrology, construction, and remote sensing. Use the improved RNPC-net model to identify the tunnel face state and extract groundwater disturbance indicators. S3. Based on elastic-plastic damage theory and seepage mechanics, a stress-seepage-heat conduction three-field coupling model is established, and the boundary conditions of the stress-seepage-heat conduction three-field coupling model are set based on the groundwater disturbance index; S4. Analyze the three-field coupling effect based on the three-field coupling model, and output the comprehensive risk level of the impact of tunnel construction on groundwater based on the comprehensive risk assessment model; S5. Use GIS spatial analysis capabilities to integrate comprehensive risk levels and combine real-time data to achieve spatial visualization and graded early warning of groundwater impacts; S6. Use the extended Kalman filter algorithm to update the geological parameters of the three-field coupling model in real time, combine it with digital twin technology to build a virtual mapping of the construction process, and search for the optimal combination of construction parameters to form a monitoring-analysis-optimization closed-loop control.

2. The method for analyzing the impact of railway tunnel construction on urban groundwater according to claim 1, characterized in that: Step S1 specifically includes the following steps: S11. Collect multi-source data: Acquire Sentinel-1A / B satellite C-band remote sensing data through EGMS to obtain surface fracture data; simultaneously acquire drilling data, geological survey reports, groundwater level monitoring data, permeability coefficient test results, and pore water pressure distribution data to form geological and hydrological data; and acquire construction process parameters and tunnel face images to form construction disturbance data; S12. Multi-source data preprocessing: For remote sensing data: First, according to the orbital geometry parameters of the Sentinel-1 satellite, the LOS displacement sequences of the ascending and descending remote sensing data are projected to the vertical direction to eliminate the radar incident angle θ deviation, and the effective monitoring points are retained through coherence threshold screening, and the corrected vertical displacement sequence d is output. vert : Where, d LOS represents the original LOS displacement time series; Then, the effective monitoring points are screened by the coherence threshold, and the principal component analysis-independent component analysis decomposition technology is used to separate the construction disturbance component and the natural settlement component in the displacement signal, and the construction disturbance displacement sequence d is extracted. excav ; For drilling data: Use screening and sedimentation methods to test particle size, obtain clay content, and then preliminarily estimate the permeability coefficient K=K0·10 based on the clay content -3 , K0 represents the baseline permeability coefficient related to the clay content, and finally the preliminarily estimated permeability coefficient K is calibrated by the water pressure test to obtain the anisotropic permeability coefficient tensor K of each layer i , where the permeability coefficient K is calculated as follows: At the same time, the degree of crack development is analyzed based on the drilling CT images, and the damage factor is defined. L fracture and L core They represent the surface crack length and core length respectively, and then ordinary Kriging interpolation is used to generate the three-dimensional damage factor field D(x, y, z).

3. The method for analyzing the impact of railway tunnel construction on urban groundwater according to claim 2, characterized in that: In step S11, the drilling data is acquired as follows: drilling is arranged every 50m-100m along the tunnel axis, and the drilling depth penetrates 10m-20m below the tunnel floor, and rock core samples are collected for lithology classification to obtain lithology distribution data; At the same time, the standard penetration test is used to obtain the number of blows N that penetrate 30 cm. SPT , combined with De Mello's empirical formula to estimate the effective internal friction angle of rock mass The construction process parameters include TBM excavation parameters and blasting construction parameters. TBM excavation parameters include cutter head torque M t , propulsion speed v p and shield pressure P shield ; Blasting construction parameters include single hole charge Q b , center of explosion distance r b , blasting time interval and surface vibration velocity v b .

4. The method for analyzing the impact of railway tunnel construction on urban groundwater according to claim 3, characterized in that: Step S2 specifically includes the following steps: S21. Constructing a multidimensional feature vector of geology, hydrology, construction, and remote sensing: S211, based on the corrected vertical displacement sequence d excav As well as the tunnel diameter R and burial depth z0, calculate the settlement characteristic parameters, which include the volume loss rate V L and slot width d; Calculation of volume loss rate V based on Gaussian sedimentation model L : Where V s represents the volume of the sedimentation tank, and ω(x) represents the vertical displacement distribution function of the surface settlement trough, and ω max represents the maximum surface settlement, ω max =max(d excav ), l i represents the distance from the inflection point of the sedimentation trough, x represents the horizontal coordinate; A t represents the tunnel cross-sectional area, and R represents the tunnel diameter; Calculate the slot width d using the following formula: Where i represents the distance from the inflection point; z0 represents the tunnel depth; S212, construct feature vector: integrate geological and hydrological data, construction data, remote sensing data and surface crack distribution data to obtain feature matrix F = [K i ,n,D,v p ,M t ,V L ,d,ω max ], n represents the surface porosity; S213, using the minimum-maximum normalization method to map each eigenvalue in the feature matrix to the interval [0,1], and obtain the normalized feature matrix F norm ; S22, Identify the tunnel face: Construct an improved RNPC-net model and use the pre-processed tunnel face image, 3D coordinates and feature matrix F norm Train the improved RNPC-net model until convergence, and output the probability distribution of weathering grade P(WD) and groundwater status P(GC): The improved RNPC-net model adopts a three-branch hybrid feature extraction module, which includes a 2D image convolution branch based on ResNet50, a 3D point cloud processing branch, and a construction parameter fully connected branch. It also fuses multi-source data inputs through an adaptive weight-assisted classifier. The cross entropy loss function is set as follows: Where L represents the loss value; y s represents the label vector; P s represents the distribution probability; S represents the total category; s represents the category index; S23, extracting groundwater disturbance indices, wherein the groundwater disturbance indices include a weathering comprehensive index WI and a groundwater activity intensity index GI; Among them, the weathering comprehensive index WI is calculated based on the weathering grade probability distribution P(WD): Where w k Represents weight; P(WD k ) represents the predicted probability of the kth weathering level, where k = 1, 2, 3, and 4 represent unweathered, slightly weathered, moderately weathered, and strongly weathered, respectively; The groundwater state probability distribution P(GC) is mapped into a numerical indicator to construct the groundwater activity intensity index GI: GI=4P(FW)+3P(DW)+2P(DP)+P(DR); Where P(FW), P(DW), P(DP), and P(DR) represent the probability of water gushing from the tunnel face, the probability of water dripping from the tunnel face, the probability of water dripping into a line on the tunnel face, and the probability of the tunnel face being dry, respectively.

5. The method for analyzing the impact of railway tunnel construction on urban groundwater according to claim 4, characterized in that: Step S3 specifically includes the following steps: S31. Elastic-Plastic Damage Stress Field Modeling: S311, inversion of damage zone parameters, including blasting damage radius and anisotropic material parameter matrix [E d ,v d ], where the blasting damage radius is estimated using the following empirical formula and calibrated with the acoustic wave test results to determine the damage area: Where R d represents the geometric parameters of the damaged area; k b represents the geological correction factor; Anisotropic material parameter matrix [E d ,v d Elastic modulus E of the damaged area in d The calculation formula is as follows: E d =e0(1-D); Where, E0 represents the elastic modulus of the original rock; Poisson's ratio v d The calculation formula is as follows: v d =v0+0.1D; Where v0 represents the Poisson's ratio of the original rock; S312, elastic zone radial effective stress σ' er and tangential effective stress σ' eθ : Where P0 represents the original rock stress; r represents the inner wall of the tunnel; R i Indicates the inner radius of the tunnel; Based on the Mohr-Coulomb criterion, the radial effective stress σ' in the plastic zone is derived. pr and tangential effective stress σ' pθ : Where, represents the cohesion in the plastic zone, and c p represents the cohesion in the plastic zone; represents the earth pressure coefficient in the plastic zone, and R p represents the radius of the plastic zone; represents the friction angle in the plastic zone, and η represents the damage reduction factor; S32. Construct a seepage equation that takes temperature effects into account and set boundary conditions to obtain the pore water pressure distribution cloud map and seepage flow rate Q. The seepage equation is expressed as follows: Where v represents the seepage velocity vector; μ represents the dynamic viscosity of water; P w represents pore water pressure; ρ represents water density; g represents gravitational acceleration; h represents water head height; β represents thermal expansion coefficient; T represents temperature field distribution; The boundary conditions are as follows: the inner wall of the tunnel r = R i Pore ​​water pressure P w =P wi , P wi Represents the water pressure in the tunnel; far field r = R w Time P w =P w0 , R w represents the seepage influence radius, P w0 represents the initial pore water pressure; The calculation formula of seepage flow Q is as follows: Where H represents the thickness of the aquifer.

6. The method for analyzing the impact of railway tunnel construction on urban groundwater according to claim 5, characterized in that: Step S4 The specific steps include: S41. Calculate the total stress by coupling the stress field and the seepage field: s r =σ′ r +P w ,s θ =s θ +P w ; Where, σ r and σ θ Represent the total radial stress and the total tangential stress respectively; σ' r and σ' θ Represent the radial stress and tangential stress of the stress field, and σ' r =σ' er +σ' pr ,σ' θ =σ' eθ + = σ′ pθ ; S42: Introduce construction heat source, calculate temperature distribution using heat conduction equation, and correct seepage parameters: Where μ(T) represents the dynamic viscosity of water after temperature correction, and μ(T)=μ0exp(-k T (T-T0)), μ0 represents the dynamic viscosity of reference water, k T represents the temperature coefficient, T and T0 represent the real-time temperature and initial temperature respectively; S43. Set the risk indicators as water inrush risk, subsidence risk and squeeze risk, among which water inrush risk Subsidence risk Squeeze risk Q th and d th are all set thresholds, r qd Indicates rock quality index; Get a single risk factor vector [RI Q ,RI d ,SPI]; S44. Constructing a comprehensive risk index (CRI): CRI=ω Q RI Q +ω d RI d +ω SPI SPI; Where, ω Q 、ω d and ω SPI Both represent weights, and ω Q +ω d +ω SPI =1; S45. Based on the comprehensive risk index (CRI), risks are divided into low risk, medium risk and high risk, and a risk-measure mapping relationship is established.

7. The method for analyzing the impact of railway tunnel construction on urban groundwater according to claim 6, characterized in that: Step S5 specifically includes the following steps: S51. Convert geological parameters, three-field coupling model analysis results and real-time monitoring data into ESRI Shapefile format and project them into the UTM coordinate system; S52. Create thematic layers of geological layers, stress fields, and seepage fields in ArcGIS Pro; S53. Use the inverse distance weighted method to interpolate discrete risk points to generate a continuous risk surface: Where R(x,y,z) represents the continuous risk value at the three-dimensional space coordinate (x,y,z); d i Represents the spatial distance between the prediction point and the i-th discrete risk monitoring point; CRI i represents the comprehensive risk index of the i-th discrete risk monitoring point; I represents the total number of discrete risk monitoring points; S54. Construct a BIM tunnel model based on tunnel geometry parameters, construction process parameters, and the analysis results of the three-field coupling model. Import the BIM tunnel model and comprehensive risk level data into the Unity 3D platform to achieve a dynamic correlation display between construction progress and risk evolution. S55. Set graded warning rules: Level 1 warning conditions: Comprehensive risk index CRI ≥ 0.7 and surface settlement rate > 1mm / day, triggering a red warning and automatically starting the emergency grouting plan; Level 2 warning conditions: 0.4≤CRI<0.7 or sedimentation rate>0.5mm / day, triggering a yellow warning and prompting increased monitoring frequency.

8. The method for analyzing the impact of railway tunnel construction on urban groundwater according to claim 7, characterized in that: Step S6 specifically includes the following steps: S61, using the extended Kalman filter to update the geological parameters of the three-field coupling model in real time, and output the updated damage factor D new and permeability coefficient K new ; Define the state vector x = [D, K i ], observation vector represents the measured value of the vertical displacement of the ground surface, Q meas Representing the measured value of seepage flow, the following state equation and observation equation are established: x k+1 =f(x k )+w k ; z k =h(x k )+v k ; Where x k+1 and x k denote the predicted value of the state vector at time k+1 and the estimated value of the state vector at time k respectively; f(·) denotes the state transfer function; w k represents state noise; z k represents the observation vector at time k; h(·) represents the observation function; v k represents the observation noise; The state estimate is iteratively calculated using the extended Kalman filter algorithm: Where, and They represent the optimal estimated value of the state vector at time k and the predicted value of the state at time k based on the k-1 time estimate; K k represents the Kalman gain, and P k|k-1 represents the covariance matrix of the k-time state prediction based on the k-1-time estimation, represents the covariance matrix of the observation noise; S62, the updated damage factor D new and permeability coefficient K new Substitute the three-field coupling model described in step S3 to recalculate the stress field, seepage field and risk index; S63, calculating the root mean square error between the weight calibration result and the measured data to evaluate the accuracy of the three-field coupling model. If the root mean square error is greater than 5%, return to step S61 and perform a secondary calibration process; S64, optimize construction parameters; S641. Build a digital twin model: Develop a digital twin of tunnel construction based on the Unity platform, integrating the BIM tunnel model and real-time monitoring data to achieve a 1:1 virtual mapping of the construction process. S642, using the proximal strategy optimization algorithm to minimize the comprehensive risk index CRI and search for the optimal parameter combination and represent the optimal support pressure and optimal excavation rate respectively.

Images