Shield tunnel underpassing highway foundation safety risk control method and system

Abstract

The application provides a shield tunnel underpassing highway foundation safety risk control method and system, relates to the shield construction risk control technical field, and includes: obtaining three-dimensional geological parameters, shield tunneling parameters, roadbed layering characteristics, historical settlement data, traffic load data and meteorological and hydrological data; modeling is carried out according to the three-dimensional geological parameters, the roadbed layering characteristics and the shield tunneling parameters, and structural response characteristic data is generated; evolution prediction is carried out according to the structural response characteristic data, the traffic load data and the historical settlement data, and a multi-stage deformation prediction field is generated; risk analysis is carried out according to the deformation prediction field and the meteorological and hydrological data, and a joint probability distribution is obtained; and a risk control strategy sequence is generated based on the joint probability distribution. The application realizes multi-stage prediction of the long-term settlement of the roadbed and the cumulative effect of the segment damage based on the cooperative evolution mechanism of the dynamic traffic load frequency and the meteorological and hydrological disturbance, and breaks through the space-time resolution limitation of the deformation field prediction under the static load assumption.

Description

Technical Field

[0001] This invention relates to the field of shield tunneling risk control technology, and more specifically, to a method and system for controlling safety risks when a shield tunnel passes under a highway subgrade. Background Technology

[0002] Currently, safety risk control for shield tunnels passing under highway subgrades is a key technical challenge in urban underground space development. The engineering field primarily relies on finite element numerical simulation methods based on geological survey data, combined with empirical safety thresholds for static risk assessment. For example, soil stress distribution is calculated using Mohr-Coulomb theory, and risk warnings are achieved by comparing monitoring data with preset thresholds. However, such methods have significant limitations in addressing the multi-field coupling effects caused by shield tunneling: complex mechanisms such as the dynamic interaction between seepage and stress fields, and the nonlinear response of the construction machinery-soil interface are difficult to model accurately, leading to significant deviations between predicted deformation values ​​and actual conditions. In recent years, although some studies have introduced machine learning algorithms for time-series analysis of settlement trends, their input features are mostly limited to single-type sensor data, failing to systematically integrate external disturbance factors such as the dynamic frequency of traffic loads and sudden changes in rainfall intensity, resulting in insufficient generalization ability of the models under complex geological conditions. Furthermore, existing risk decision-making is mostly based on fixed rule bases, lacking the ability to dynamically optimize parameters driven by real-time monitoring data.

[0003] Based on the shortcomings of the existing technologies, there is an urgent need for a method and system for controlling safety risks when shield tunnels pass under highway subgrades. Summary of the Invention

[0004] The purpose of this invention is to provide a method and system for controlling safety risks when a shield tunnel passes under a highway subgrade, in order to improve the aforementioned problems. To achieve the above objective, the technical solution adopted by this invention is as follows:

[0005] Firstly, this application provides a method for controlling safety risks when a shield tunnel passes under a highway subgrade, including:

[0006] Acquire three-dimensional geological parameters, shield tunneling parameters, subgrade layering characteristics, historical settlement data, traffic load data, and meteorological and hydrological data;

[0007] Based on the three-dimensional geological parameters, the subgrade layering characteristics, and the shield tunneling parameters, a geological-shield-subgrade interaction model is constructed, and structural response characteristic data is generated.

[0008] Evolution prediction is performed based on the structural response characteristic data, the traffic load data, and the historical settlement data to generate a multi-stage deformation prediction field.

[0009] Risk analysis is performed based on the deformation prediction field and the meteorological and hydrological data to calculate the joint probability distribution.

[0010] Based on the joint probability distribution and real-time monitoring data, a risk control strategy sequence is generated.

[0011] Secondly, this application also provides a safety risk control system for shield tunnels passing under highway subgrades, including:

[0012] The acquisition module is used to acquire three-dimensional geological parameters, shield tunneling parameters, roadbed layering characteristics, historical settlement data, traffic load data, and meteorological and hydrological data.

[0013] The modeling module is used to perform modeling processing based on the three-dimensional geological parameters, the subgrade layering characteristics, and the shield tunneling parameters, to construct a geological-shield-subgrade interaction model, and to generate structural response feature data.

[0014] The prediction module is used to predict the evolution based on the structural response characteristic data, the traffic load data and the historical settlement data, and generate a multi-stage deformation prediction field.

[0015] The analysis module is used to perform risk analysis based on the deformation prediction field and the meteorological and hydrological data, and calculate the joint probability distribution.

[0016] The generation module generates a risk control strategy sequence based on the joint probability distribution and real-time monitoring data.

[0017] The beneficial effects of this invention are as follows:

[0018] This invention significantly improves the calculation accuracy of the relationship between soil seepage-stress interaction and segment deformation coordination by constructing a multi-physics dynamic coupling model of geology-shield-subgrade, overcoming the interface response distortion problem caused by separate modeling in traditional methods. Based on the co-evolution mechanism of dynamic traffic load frequency and meteorological and hydrological disturbances, it realizes multi-stage prediction of the cumulative effect of long-term subgrade settlement and segment damage, breaking through the spatiotemporal resolution limitations of deformation field prediction under static load assumptions. By integrating Bayesian networks and risk chain propagation algorithms, it effectively reduces the assessment bias of multi-factor coupled risk events. Attached Figure Description

[0019] To more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings used in the embodiments will be briefly introduced below. It should be understood that the following drawings only show some embodiments of the present invention and should not be regarded as a limitation on the scope. For those skilled in the art, other related drawings can be obtained based on these drawings without creative effort.

[0020] Figure 1This is a schematic diagram of a method for controlling safety risks when a shield tunnel passes under a highway subgrade, as described in an embodiment of the present invention.

[0021] Figure 2 This is a schematic diagram of a safety risk control system for a shield tunnel passing under a highway subgrade, as described in an embodiment of the present invention.

[0022] Figure 3 This is a schematic diagram of a safety risk control device for a shield tunnel passing under a highway subgrade, as described in an embodiment of the present invention.

[0023] The diagram is labeled as follows: 800, a safety risk control device for shield tunnels passing under highway subgrades; 801, processor; 802, memory; 803, multimedia component; 804, I / O interface; 805, communication component; 901, acquisition module; 902, modeling module; 903, prediction module; 904, analysis module; 905, generation module. Detailed Implementation

[0024] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, 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, not all, of the embodiments of the present invention. The components of the embodiments of the present invention described and shown in the accompanying drawings can generally be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of the present invention provided in the accompanying drawings is not intended to limit the scope of the claimed invention, but merely to illustrate selected embodiments of the invention. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without inventive effort are within the scope of protection of the present invention.

[0025] It should be noted that similar reference numerals and letters in the following figures indicate similar items; therefore, once an item is defined in one figure, it does not need to be further defined and explained in subsequent figures. Furthermore, in the description of this invention, terms such as "first," "second," etc., are used only to distinguish descriptions and should not be construed as indicating or implying relative importance.

[0026] Example 1:

[0027] This embodiment provides a method for controlling safety risks when a shield tunnel passes under a highway subgrade.

[0028] See Figure 1 The figure shows that the method includes steps S100 to S500.

[0029] Step S100: Obtain three-dimensional geological parameters, shield tunneling parameters, subgrade layering characteristics, historical settlement data, traffic load data, and meteorological and hydrological data;

[0030] Understandably, the three-dimensional geological parameters encompass the spatial distribution characteristics of soil permeability coefficient, elastic modulus, and stratum interface dip angle, obtained through the fusion modeling of ground-penetrating radar and borehole data; the shield tunneling parameters include dynamic fluctuations in cutterhead torque, propulsion speed gradient, and time-series data of soil chamber pressure, recorded in real time by the shield machine's PLC system; the subgrade layering characteristics are obtained through in-situ shear wave velocity testing and geotechnical tests to obtain the stiffness anisotropy matrix of each soil layer, used to represent the mechanical response characteristics of the embankment subgrade; historical settlement data integrates long-term monitoring curves from shield tunneling projects in similar strata, used to reveal the creep law of the soil; traffic load data is described using dynamic axle load spectrum and traffic flow density distribution function, generated through real-time traffic flow data collected by the highway ETC system; meteorological and hydrological data include spatiotemporal distribution data of regional rainfall intensity and groundwater level fluctuation monitoring values.

[0031] Step S200: Based on the three-dimensional geological parameters, subgrade layering characteristics and shield tunneling parameters, modeling is performed to construct a geological-shield-subgrade interaction model and generate structural response characteristic data;

[0032] It should be noted that this step constructs a heterogeneous geological model reflecting the actual geological structure by using three-dimensional geological parameters and the layered characteristics of the roadbed. Combined with dynamic construction data from the tunnel boring machine (TBM) parameters, such as changes in cutterhead rotation force and propulsion speed fluctuations, a dynamic interaction relationship between the construction machinery and the geological body is established. In this process, a numerical calculation method coupling geological conditions and construction parameters is employed to jointly calculate changes in soil internal stress, TBM contact pressure, and the influence of seepage water pressure, generating structural response characteristic data including key indicators such as soil stress concentration zones, segment deformation, and interface slippage trends. This step integrates static geological properties, dynamic construction disturbances, and structural mechanical responses into a unified model. Through real-time parameter updates, it simulates the dynamic changes of soil and structure during construction, solving the problem of model deviation from actual working conditions caused by fixed parameter assumptions in traditional methods. This provides a high-precision input basis for subsequent risk prediction.

[0033] Step S300: Based on structural response characteristic data, traffic load data and historical settlement data, perform evolution prediction and generate a multi-stage deformation prediction field;

[0034] Understandably, this step, based on the distribution of soil stress concentration zones and segment deformation trends in structural response characteristic data, and combined with the dynamic fluctuation characteristics of traffic flow density in traffic load data (such as the surge in axle load frequency during peak hours), uses a dynamic mode decomposition algorithm to separate the transient vibration component (caused by sudden traffic loads) and the steady-state creep component (formed by long-term load accumulation). Furthermore, it integrates the settlement rate decay law under similar geological conditions in historical settlement data, calibrates the time-varying characteristics of materials through parameter adaptive inversion technology, and constructs a multi-stage prediction model considering the synergistic effect of traffic load cycles and soil stiffness degradation. Preferably, in this process, the initial damage field caused by shield tunneling disturbance is embedded as a boundary condition into the prediction model. A time convolutional network is used to capture the nonlinear correlation between pore water pressure dissipation and structural deformation accumulation, ultimately generating a graded deformation prediction field that includes the short-term construction disturbance stage, the medium-term traffic load stage, and the long-term environmental factor influence stage.

[0035] Step S400: Conduct risk analysis based on the deformation prediction field and meteorological and hydrological data, and calculate the joint probability distribution;

[0036] Specifically, firstly, based on the dynamic changes in rainfall intensity and groundwater level in meteorological data, a time-varying model of rainfall permeability coefficient is established to quantify the enhancing effect of pore water pressure on soil softening. Secondly, combining the segment bending moment gradient distribution and subgrade settlement rate in the deformation prediction field, a spatially variable risk index matrix is ​​generated using random field theory. Thirdly, through Bayesian network topology construction, geological vulnerability parameters, deformation evolution trends, and external environmental disturbances are probabilistically correlated, and Monte Carlo simulation is used to calculate the joint triggering probability of multiple risk events (subgrade instability, segment cracking, and surface collapse) in the spatiotemporal dimensions.

[0037] Step S500: Generate a risk control strategy sequence based on the joint probability distribution and real-time monitoring data.

[0038] It should be noted that this step, based on the spatiotemporal triggering probability of different risk events in the joint probability distribution, and combined with key indicators such as real-time monitoring of settlement rate and pore water pressure, constructs a multi-objective optimization model that includes construction safety constraints and structural damage thresholds. Risk probabilities are transformed into dynamic boundary conditions for control parameters such as tunneling speed and soil chamber pressure. A set of candidate strategies that balances risk suppression and construction efficiency is generated through non-convex optimization. Furthermore, a reinforcement learning mechanism driven by real-time monitoring data streams is introduced, utilizing a strategy evaluation-feedback loop to continuously optimize parameter priorities, ultimately outputting a sequence of tunneling control commands that adjusts in real-time according to geological response.

[0039] Further, step S200 includes steps S210 to S230.

[0040] Step S210: Model the geological-subgrade interface based on the three-dimensional geological parameters and subgrade layering characteristics to obtain the coupled stress field;

[0041] In this step, the geological-subgrade interface modeling integrates three-dimensional geological parameters (such as soil hardness and spatial distribution of permeability) with subgrade layering characteristics (such as differences in compressibility and shear strength of soil at different depths), and constructs a heterogeneous stratification model using a layered integral algorithm. Based on the anisotropic parameters of soil permeability coefficient and stiffness, the interaction force between groundwater level fluctuations and subgrade fill materials is calculated using the elastoplastic theory of porous media, generating a stress field distribution reflecting the water-soil coupling effect, thus overcoming the limitation of traditional homogeneous models in simplifying the interface mechanical response.

[0042] Step S220: Model the shield-soil contact based on the shield tunneling parameters and coupled stress field. Generate a dynamic contact stress field by fusing the dynamic parameters of the cutterhead torque with the characteristics of the propulsion speed gradient.

[0043] Understandably, shield-soil contact modeling inputs dynamic parameters of the cutterhead torque (such as torque fluctuation spectrum characteristics) and spatial gradient changes in the propulsion speed (such as differences in propulsion rates in different regions of the cutterhead) into the contact dynamics equations to solve for the nonlinear frictional boundary conditions at the contact surface between the shield shell and the soil. By updating the matching relationship between the shield tunneling direction and the soil stress state in real time, a dynamically changing contact pressure field is constructed to accurately characterize the instantaneous disturbance effect of the construction machinery on the soil.

[0044] Step S230: Perform multi-field coupling based on the coupled stress field and the dynamic contact stress field, and obtain structural response characteristic data by jointly calculating the seepage pressure gradient, soil stress redistribution and segment deformation coordination relationship.

[0045] It should be noted that the multi-field coupling processing uses a fluid-structure interaction iterative algorithm to bidirectionally couple the seepage pressure gradient (calculated using a modified Darcy's law formula) with the soil stress redistribution (based on dynamic contact stress field updates), while simultaneously introducing coordination constraints between the segment stiffness matrix and the soil displacement field. An explicit dynamic algorithm is used to simultaneously solve the interaction between seepage, stress, and structural deformation, outputting comprehensive response data including the expansion of the soil plastic zone, the segment bending moment anomaly zone, and the seepage instability risk zone, thus achieving collaborative simulation of multi-physics fields under construction disturbances.

[0046] The formula for calculating structural response characteristic data is:

[0047]

[0048] F=∫ Ω (σ soil -σ shield )·δ u dΩ=0;

[0049] Where, σ total σ represents the total stress tensor; soil σ represents the soil stress tensor; shield This represents the contact stress tensor of the tunnel boring machine. Represents the seepage pressure gradient; C represents the seepage-stress coupling coefficient matrix; K represents the segment stiffness matrix; u pipe Represents the segment displacement field; F represents the compatibility constraint equation; δ u Ω represents the virtual displacement field; d represents the computational domain; and d represents the differential.

[0050] Further, step S300 includes steps S310 to S330.

[0051] Step S310: Decouple the dynamic response based on the structural response characteristic data and traffic load data, and generate a dynamic stress-strain evolution sequence by separating the transient vibration and steady-state creep components of the soil.

[0052] Understandably, this step addresses the combined effects of construction disturbance and traffic load by employing a dynamic decomposition method to extract transient vibration (such as high-frequency soil fluctuations caused by sudden heavy loads) and steady-state creep (such as low-frequency cumulative deformation caused by long-term traffic flow) components from structural response data, thereby constructing an evolutionary sequence reflecting mechanical behavior at different time scales.

[0053] Step S320: Perform parameter inversion based on the dynamic stress-strain evolution sequence and historical settlement data to obtain the time-varying constitutive equation of the material;

[0054] Specifically, this step addresses the dynamic modeling problem of material properties changing with time and environment. Based on the transient and steady-state components of the dynamic stress-strain evolution sequence, combined with the long-term deformation attenuation patterns recorded in historical settlement data, a particle swarm optimization algorithm is used to inversely solve for the viscoelastic parameters of the soil. The specific process includes: constructing a generalized Burgers creep equation incorporating the pore water pressure dissipation rate; using the transient vibration component as the initial boundary condition and the steady-state creep component as the objective optimization function; and iteratively adjusting the damage factor and seepage coupling term in the equation to minimize the fitting error between the model output and the historical settlement curve. The resulting time-varying constitutive equation not only includes the viscoelastic-plastic parameters of the traditional creep model but also incorporates rainfall intensity and groundwater level fluctuations as external disturbance variables. The constitutive equation takes the following form:

[0055]

[0056] Where σ(t) is the time-dependent stress tensor, representing the stress state of the material at time t; ∈(t) represents the deformation response of the material under external load, including elastic deformation and viscous flow components. E1 represents the rate of change of strain with time; η1 represents the transient elastic modulus; E2 represents the energy dissipation characteristics of the material during transient deformation; η2 represents the steady-state viscoelastic modulus; η2 represents the steady-state viscosity coefficient; and τ represents the integral variable.

[0057] Step S330: Based on the time-varying constitutive equation of the material and the dynamic stress-strain evolution sequence, a multi-stage deformation prediction field is generated by simulating the synergistic effect of soil pore water pressure dissipation and structural stiffness degradation under traffic load cycle.

[0058] Specifically, based on the dynamic parameters in the time-varying constitutive equations of materials and the load spectrum characteristics in the dynamic stress-strain evolution sequence, this step employs a time-convolutional network to construct a correlation model between the number of traffic load cycles and the accumulation of soil damage. For example, daily traffic flow data is mapped to equivalent cyclic loading times, and the expanded convolutional layers of the time-convolutional network capture the synergistic effect of pore water pressure dissipation (such as changes in consolidation rate caused by pore water discharge after each traffic vibration) and soil stiffness degradation (such as microcrack propagation caused by each loading). The resulting multi-stage deformation prediction field comprises three core stages: segment installation misalignment and initial soil consolidation during the construction disturbance-dominated period (0-3 months); cumulative settlement and crack propagation during the traffic load period (3-24 months); and groundwater chemical corrosion and material aging effects during the long-term environmental evolution period (>24 months).

[0059] Further, step S400 includes steps S410 to S430.

[0060] Step S410: Based on the deformation prediction field and meteorological and hydrological data, seepage-deformation coupling is performed. By quantifying the enhancing effect of pore water pressure on soil softening, a multi-field coupled deformation field is generated.

[0061] Understandably, this step employs unsaturated soil mechanics theory, converting rainfall intensity into a spatiotemporal variation function of soil moisture content, and calculating the weakening effect of seepage pressure gradient on effective soil stress using a modified Darcy's law formula. Simultaneously, by combining the initial stress state in the deformation prediction field (such as soil stress release caused by tunnel boring), the bidirectional coupling relationship between the seepage field and stress field is iteratively updated, quantifying the amplifying effect of soil softening caused by continuous rainfall on cumulative deformation. The resulting multi-field coupled deformation field includes the displacement increment distribution corrected for seepage pressure and the identification of soil strength deterioration hotspots, providing accurate input for risk factor correlation.

[0062] Step S420: Correlate risk factors based on multi-field coupled deformation fields and three-dimensional geological parameters, and generate a conditional probability relationship diagram by constructing a Bayesian network topology.

[0063] Understandably, this step, based on key indicators in the multi-field coupled deformation field (such as the extreme value of segment bending moment and subgrade settlement curvature) and three-dimensional geological parameters (such as the spatial distribution of soil permeability coefficient and the anisotropy of stiffness in the embankment area), adopts a hybrid-driven Bayesian network construction method: First, risk-sensitive units such as soil permeability anomaly zones and weak interlayer distribution are generated through a geological parameter random field model; second, the PC algorithm (Peter-Clark causal discovery) is used to extract statistical correlations between events such as abrupt changes in subgrade settlement rate and stress concentration of segments from the deformation field data; finally, the network topology is constrained by engineering prior knowledge to construct a three-layer directed acyclic graph containing root nodes (geological vulnerability indicators), intermediate nodes (deformation evolution characteristics), and leaf nodes (risk events). In the network parameter learning stage, the Bayesian Markov chain Monte Carlo method is used to fuse historical disaster case data to determine the conditional probability table, and finally outputs a risk conditional probability relationship graph with spatial variability.

[0064] Step S430: Perform risk propagation calculation based on the conditional probability relationship diagram. By integrating the seepage mutation mechanism triggered by rainfall intensity threshold and the cumulative damage effect associated with traffic load frequency, generate a spatiotemporal joint probability distribution matrix of roadbed instability, segment cracking and surface collapse.

[0065] It should be noted that this step first maps the seepage mutation mechanism triggered by rainfall intensity thresholds to the probability update event of the root node in a Bayesian network; secondly, it dynamically adjusts the conditional probability distribution of intermediate nodes through a traffic load frequency-cumulative damage conversion model; finally, it uses Monte Carlo simulation to sample in parallel across the spatiotemporal dimensions to calculate the joint probability distribution of roadbed instability, segment cracking, and surface collapse. Specifically, risk units are divided spatially based on random fields of geological parameters, and seasonal periodic characteristics of meteorological data are embedded in the temporal dimension. The generated three-dimensional joint probability matrix (spatial grid-time window-risk type) can accurately characterize the concurrent probability of multiple risks in a specific area within a specified time period, providing a quantitative basis for dynamic risk management.

[0066] Formula for calculating the joint probability distribution matrix:

[0067]

[0068] Among them, P RPS (x,t) represents the three-dimensional joint probability matrix, with the dimensions corresponding to the risk probabilities of roadbed instability (R), segment cracking (P), and surface collapse (S), respectively; Φ(·) represents the seepage mutation probability function; I(x,t) represents the rainfall intensity at spatial location x at time t; θ represents the seepage mutation trigger threshold; δ represents the scale parameter of the seepage mutation probability. Γ(x) represents the Hadamard product; η represents the cumulative damage attenuation coefficient of traffic load; N(x,t) represents the traffic load frequency at location x at time t; Γ(x) represents the spatial geological vulnerability distribution matrix; Λ(t) represents the time attenuation factor.

[0069] Further, step S420 includes steps S421 to S423.

[0070] Step S421: Extract risk indicators based on multi-field coupled deformation field and three-dimensional geological parameters to obtain risk feature vectors reflecting structural damage sensitivity;

[0071] Specifically, this step spatially discretizes the deformation field data, calculating the covariance matrix of deformation indices (such as the standard deviation of segment bending moment gradient and the second derivative of settlement curvature) and geological parameters (such as the coefficient of variation of permeability coefficient and the decay rate of fill modulus) within each cell. Principal component analysis is then used to reduce the dimensionality and extract the first three principal components, forming a three-dimensional risk feature vector that includes the segment stress concentration sensitivity, the tendency of subgrade seepage instability, and the rate of soil stiffness degradation. This vector quantifies the spatial heterogeneous response characteristics of structural damage under multi-physics coupling.

[0072] Step S422: Perform causal topology modeling based on the risk feature vector, and use the physical correlation rules between the deterioration of the subgrade fill modulus and the stress concentration of the tunnel segments to generate an initial directed acyclic graph containing the causal relationship between settlement exceeding the limit and tunnel segment cracking.

[0073] In this step, the mechanical transmission mechanism between the modulus degradation of the subgrade fill (stiffness degradation rate component in the eigenvector) and the stress concentration of the tunnel segments (stress sensitivity component in the eigenvector) is utilized (e.g., the decrease in modulus leads to the redistribution of soil stress and induces additional bending moments in the tunnel segments). A constrained structural learning algorithm is used to generate an initial directed acyclic graph.

[0074] Step S423: Perform dynamic probability learning based on the initial directed acyclic graph, and output a Bayesian network topology graph by quantifying the conditional dependence strength of risk events within different geological units.

[0075] Specifically, this step, based on the causal topology defined in the initial directed acyclic graph, divides the study area into heterogeneous geological units according to three-dimensional geological parameters. Within each unit, historical disaster case data and real-time monitored deformation characteristics are used to learn a conditional probability table through Markov chain Monte Carlo sampling. A dynamic weight allocation mechanism is introduced to dynamically adjust the parameter update frequency based on real-time monitoring data, ensuring that network parameters adaptively evolve with environmental disturbances. The final output Bayesian network topology graph contains spatially embedded conditional probability distributions (such as contour maps of the probability of segment cracking within different geological units), achieving refined and localizable risk quantification assessment and providing a probabilistic reasoning basis for dynamic decision-making.

[0076] Further, step S500 includes steps S510 to S530.

[0077] Step S510: Perform dynamic risk mapping based on the joint probability distribution and real-time monitoring data. By fusing the displacement difference rate of monitoring points with the pore water pressure mutation threshold, generate a real-time corrected field of the spatiotemporal risk probability density function.

[0078] Preferably, the correction method includes: mapping the displacement difference rate to a spatial risk heterogeneity weight, recalibrating the local probability distribution triggered by a sudden change in pore water pressure, and finally generating a risk correction field covering the construction area with a time resolution of minutes, thereby realizing the dynamic tracking and quantification of risk hotspots.

[0079] Step S520: Optimize based on real-time correction field and shield tunneling parameters by solving a non-convex optimization problem that includes segment damage tolerance constraints and surface settlement gradient constraints to generate a Pareto front solution set.

[0080] It should be noted that the objective function includes minimizing the risk probability, controlling the surface settlement gradient, and maximizing tunneling efficiency. Constraints cover segment damage tolerance and equipment safety thresholds. Furthermore, the NSGA-III algorithm (a non-dominated genetic algorithm based on reference points) is used to solve this non-convex optimization problem: an adaptive reference point generation strategy is used to process the high-dimensional objective space, constraint dominance relations are introduced to screen feasible solutions, and finally, a Pareto front solution set is output, containing 50 sets of non-dominated solutions for tunneling speed, grouting volume, and soil chamber pressure, providing diverse choices for strategy decision-making.

[0081] Step S530: Verify based on the Pareto front solution set, and select the optimal strategy sequence that meets the real-time risk threshold by dynamically adjusting the priority of control parameters.

[0082] Understandably, this step uses candidate strategies from the Pareto solution set, employs Monte Carlo random perturbation simulation to evaluate strategy robustness, and calculates the risk threshold deviation, such as the probability that the actual settlement exceeds the predicted value by less than 5%. Combined with real-time monitoring of pore water pressure early warning levels, the priority of control parameters is dynamically adjusted using a Q-learning reinforcement learning model. Ultimately, the strategy sequence that satisfies real-time risk constraints and has the optimal construction efficiency is selected, forming a closed-loop control command that is then issued to the tunnel boring machine control system.

[0083] Further, step S530 includes steps S531 to S533.

[0084] Step S531: Perform parameter sensitivity analysis based on the Pareto front solution set, and extract the combination of dominant control parameters by combining the shield soil chamber pressure threshold and grouting rate boundary conditions.

[0085] It should be noted that this step identifies key decision variables by quantifying the influence weights of control parameters in the Pareto solution set on risk and efficiency indicators. Based on the parameter combination set generated by multi-objective optimization, the dimensions of the principal control parameters that significantly contribute to the risk variance are extracted. Combined with the soil pressure threshold and grouting rate boundary in the engineering constraints, the dominant parameter combinations are selected. By revealing the differences in the sensitivity of multivariate coupling effects to risk suppression, the core role of the synergistic regulation of grouting rate and soil pressure in settlement control is clarified.

[0086] Step S532: Based on the combination of dominant control parameters and real-time monitoring data, the robustness of the strategy is verified. By integrating Monte Carlo random disturbance simulation with real-time risk threshold deviation assessment, a subset of candidate strategies is selected in the tunneling speed-grouting volume parameter space.

[0087] Understandably, this step involves fusing Monte Carlo random perturbation simulations with real-time monitoring data to assess the adaptability of the dominant parameter combination under complex working conditions. Perturbation scenarios such as grouting failure and sensor drift are constructed within the tunneling speed-grouting volume parameter space to quantify the deviation of risk thresholds during strategy execution. The assessment weights are then dynamically adjusted based on real-time monitoring of cutterhead torque fluctuations and pore water pressure change rates.

[0088] Step S533: Dynamically adjust the priority based on the candidate policy subset, dynamically update the parameter weights based on the online decision tree generation mechanism of reinforcement learning, and output the optimal policy sequence.

[0089] This step utilizes reinforcement learning to achieve adaptive strategy optimization. Using real-time monitoring data as input, a Q-learning model is constructed to dynamically adjust control parameter weights. An online decision tree generation mechanism integrates historical strategy execution results with current working condition characteristics, iteratively updating the priorities of parameters such as grouting volume and tunneling speed. Finally, a phased, progressive optimal strategy sequence is output, forming a closed-loop control command flow that balances dynamic risk management and construction efficiency, driving the tunnel boring machine control system in real-time to achieve adaptive tunneling.

[0090] Example 2:

[0091] like Figure 2 As shown in the figure, this embodiment provides a safety risk control system for shield tunnels passing under highway subgrades. The system includes:

[0092] The acquisition module 901 is used to acquire three-dimensional geological parameters, shield tunneling parameters, roadbed layering characteristics, historical settlement data, traffic load data, and meteorological and hydrological data.

[0093] Modeling module 902 is used to perform modeling processing based on three-dimensional geological parameters, subgrade layering characteristics and shield tunneling parameters, construct a geological-shield-subgrade interactive model, and generate structural response characteristic data;

[0094] Prediction module 903 is used to predict the evolution based on structural response characteristic data, traffic load data and historical settlement data, and generate a multi-stage deformation prediction field.

[0095] Analysis module 904 is used to perform risk analysis based on deformation prediction field and meteorological and hydrological data, and calculate the joint probability distribution;

[0096] The generation module 905 generates a risk control strategy sequence based on the joint probability distribution and real-time monitoring data.

[0097] In one specific embodiment of the present invention, the modeling module 902 includes:

[0098] The first modeling unit is used to model the geological-subgrade interface based on three-dimensional geological parameters and subgrade layering characteristics to obtain the coupled stress field.

[0099] The second modeling unit is used to model the shield-soil contact based on the shield tunneling parameters and coupled stress field. It generates a dynamic contact stress field by integrating the dynamic parameters of the cutterhead torque and the characteristics of the propulsion speed gradient.

[0100] The third modeling unit is used to perform multi-field coupling based on the coupled stress field and the dynamic contact stress field, and obtain structural response characteristic data by jointly calculating the seepage pressure gradient, soil stress redistribution and segment deformation coordination relationship.

[0101] In one specific embodiment of the present invention, the prediction module 903 includes:

[0102] The first prediction unit is used to decouple the dynamic response based on the structural response characteristic data and traffic load data. By separating the transient vibration and steady-state creep components of the soil, a dynamic stress-strain evolution sequence is generated.

[0103] The second prediction unit is used to perform parameter inversion based on the dynamic stress-strain evolution sequence and historical settlement data to obtain the time-varying constitutive equation of the material.

[0104] The third prediction unit is used to make predictions based on the material time-varying constitutive equation and dynamic stress-strain evolution sequence. It generates a multi-stage deformation prediction field by simulating the synergistic effect of soil pore water pressure dissipation and structural stiffness degradation under traffic load cycles.

[0105] Example 3:

[0106] Corresponding to the above method embodiments, this embodiment also provides a safety risk control device for shield tunnels passing under highway subgrades. The safety risk control device for shield tunnels passing under highway subgrades described below can be referred to in correspondence with the safety risk control method for shield tunnels passing under highway subgrades described above.

[0107] Figure 3 This is a block diagram illustrating a safety risk control device 800 for a shield tunnel passing under a highway subgrade, according to an exemplary embodiment. Figure 3 As shown, the shield tunnel under highway subgrade safety risk control device 800 may include: a processor 801 and a memory 802. The shield tunnel under highway subgrade safety risk control device 800 may also include one or more of the following: a multimedia component 803, an I / O interface 804, and a communication component 805.

[0108] The processor 801 controls the overall operation of the shield tunnel under highway subgrade safety risk control device 800 to complete all or part of the steps in the aforementioned shield tunnel under highway subgrade safety risk control method. The memory 802 stores various types of data to support the operation of the shield tunnel under highway subgrade safety risk control device 800. This data may include, for example, instructions for any application or method operating on the shield tunnel under highway subgrade safety risk control device 800, as well as application-related data such as contact data, sent and received messages, images, audio, video, etc. The memory 802 can be implemented using any type of volatile or non-volatile storage device or a combination thereof, such as Static Random Access Memory (SRAM), Electrically Erasable Programmable Read-Only Memory (EEPROM), Erasable Programmable Read-Only Memory (EPROM), Programmable Read-Only Memory (PROM), Read-Only Memory (ROM), magnetic storage, flash memory, magnetic disk, or optical disk. The multimedia component 803 may include a screen and an audio component. The screen may be, for example, a touchscreen, and the audio component is used to output and / or input audio signals. For example, the audio component may include a microphone for receiving external audio signals. The received audio signals may be further stored in the memory 802 or transmitted via the communication component 805. The audio component also includes at least one speaker for outputting audio signals. I / O interface 804 provides an interface between processor 801 and other interface modules, such as keyboards, mice, and buttons. These buttons can be virtual or physical. Communication component 805 is used for wired or wireless communication between the shield tunnel under highway subgrade safety risk control device 800 and other devices. Wireless communication includes, for example, Wi-Fi, Bluetooth, Near Field Communication (NFC), 2G, 3G, or 4G, or a combination thereof. Therefore, the corresponding communication component 805 may include: a Wi-Fi module, a Bluetooth module, and an NFC module.

[0109] In an exemplary embodiment, a safety risk control device 800 for a shield tunnel passing under a highway subgrade can be implemented by one or more application-specific integrated circuits (ASICs), digital signal processors (DSPs), digital signal processing devices (DSPDs), programmable logic devices (PLDs), field programmable gate arrays (FPGAs), controllers, microcontrollers, microprocessors, or other electronic components to perform the aforementioned safety risk control method for a shield tunnel passing under a highway subgrade.

[0110] In another exemplary embodiment, a computer-readable storage medium including program instructions is also provided. When executed by a processor, these program instructions implement the steps of the aforementioned method for controlling the safety risks of a shield tunnel passing under a highway subgrade. For example, the computer-readable storage medium may be the aforementioned memory 802 including program instructions, which may be executed by a processor 801 of a shield tunnel under a highway subgrade safety risk control device 800 to complete the aforementioned method for controlling the safety risks of a shield tunnel passing under a highway subgrade.

[0111] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any changes or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in the present invention should be included within the scope of protection of the present invention.

Claims

1. A method for controlling safety risks when a shield tunnel passes under a highway subgrade, characterized in that, include: Acquire three-dimensional geological parameters, shield tunneling parameters, subgrade layering characteristics, historical settlement data, traffic load data, and meteorological and hydrological data; Based on the three-dimensional geological parameters, the subgrade layering characteristics, and the shield tunneling parameters, a geological-shield-subgrade interaction model is constructed, and structural response characteristic data is generated. Evolution prediction is performed based on the structural response characteristic data, the traffic load data, and the historical settlement data to generate a multi-stage deformation prediction field. Risk analysis is performed based on the deformation prediction field and the meteorological and hydrological data to calculate the joint probability distribution. Based on the joint probability distribution and real-time monitoring data, a risk control strategy sequence is generated; The risk analysis based on the deformation prediction field and the meteorological and hydrological data includes: Based on the deformation prediction field and meteorological and hydrological data, seepage-deformation coupling is performed, and a multi-field coupled deformation field is generated by quantifying the enhancing effect of pore water pressure on soil softening. Risk factors are correlated based on the multi-field coupled deformation field and the three-dimensional geological parameters, and a conditional probability relationship diagram is generated by constructing a Bayesian network topology. Risk propagation calculations are performed based on the conditional probability relationship diagram. By integrating the seepage mutation mechanism triggered by rainfall intensity threshold and the cumulative damage effect associated with traffic load frequency, a spatiotemporal joint probability distribution matrix of roadbed instability, segment cracking and surface collapse is generated. The risk factor correlation based on the multi-field coupled deformation field and the three-dimensional geological parameters includes: Risk indicators are extracted based on the multi-field coupled deformation field and the three-dimensional geological parameters to obtain a risk feature vector reflecting the structural damage sensitivity. Based on the risk feature vector, causal topology modeling is performed, and an initial directed acyclic graph containing the causal relationship between excessive settlement and segment cracking is generated by utilizing the physical correlation rules between the deterioration of the subgrade fill modulus and the stress concentration of the tunnel segments. Dynamic probability learning is performed based on the initial directed acyclic graph. By quantifying the conditional dependence strength of risk events within different geological units, a Bayesian network topology graph is output.

2. The method for controlling safety risks when a shield tunnel passes under a highway subgrade according to claim 1, characterized in that, Modeling is performed based on the three-dimensional geological parameters, the subgrade layering characteristics, and the shield tunneling parameters, including: Based on the three-dimensional geological parameters and the subgrade layering characteristics, a geological-subgrade interface model is performed to obtain the coupled stress field. Based on the shield tunneling parameters and the coupled stress field, shield-soil contact modeling is performed, and a dynamic contact stress field is generated by fusing the dynamic parameters of cutterhead torque and the characteristics of propulsion speed gradient. By coupling the coupled stress field with the dynamic contact stress field in multiple fields, and by jointly calculating the seepage pressure gradient, soil stress redistribution, and segment deformation coordination relationship, structural response characteristic data are obtained.

3. The method for controlling safety risks when a shield tunnel passes under a highway subgrade according to claim 1, characterized in that, Evolution prediction is performed based on the structural response characteristic data, the traffic load data, and the historical settlement data, including: Dynamic response decoupling is performed based on the structural response characteristic data and the traffic load data. By separating the transient vibration and steady-state creep components of the soil, a dynamic stress-strain evolution sequence is generated. Based on the dynamic stress-strain evolution sequence and the historical settlement data, parameter inversion is performed to obtain the time-varying constitutive equation of the material; Based on the time-varying constitutive equation of the material and the dynamic stress-strain evolution sequence, a multi-stage deformation prediction field is generated by simulating the synergistic effect of soil pore water pressure dissipation and structural stiffness degradation under traffic load cycles.

4. The method for controlling safety risks when a shield tunnel passes under a highway subgrade according to claim 1, characterized in that, Based on the joint probability distribution and real-time monitoring data, a risk control strategy sequence is generated, including: Dynamic risk mapping is performed based on the joint probability distribution and real-time monitoring data. By fusing the displacement difference rate of monitoring points with the pore water pressure mutation threshold, a real-time correction field of the spatiotemporal risk probability density function is generated. Based on the real-time correction field and the shield tunneling parameters, optimization is performed. By solving a non-convex optimization problem that includes segment damage tolerance constraints and surface settlement gradient constraints, a Pareto front solution set is generated. The optimal strategy sequence that meets the real-time risk threshold is verified by dynamically adjusting the priority of control parameters and selecting the solution set based on the Pareto front.

5. The method for controlling safety risks when a shield tunnel passes under a highway subgrade according to claim 4, characterized in that, Verification is performed based on the Pareto front solution set, including: Based on the Pareto front solution set, parameter sensitivity analysis is performed, and the combination of dominant control parameters is extracted by combining the boundary conditions of the shield tunnel soil pressure threshold and the grouting rate. Based on the combination of dominant control parameters and the real-time monitoring data, the robustness of the strategy is verified. By integrating Monte Carlo random disturbance simulation with real-time risk threshold deviation assessment, a subset of candidate strategies is selected in the tunneling speed-grouting volume parameter space. Dynamic priority adjustment is performed based on the candidate policy subset, and the parameter weights are dynamically updated and the optimal policy sequence is output based on the online decision tree generation mechanism of reinforcement learning.

6. A safety risk control system for shield tunnels passing under highway subgrades, characterized in that, include: The acquisition module is used to acquire three-dimensional geological parameters, shield tunneling parameters, roadbed layering characteristics, historical settlement data, traffic load data, and meteorological and hydrological data. The modeling module is used to perform modeling processing based on the three-dimensional geological parameters, the subgrade layering characteristics, and the shield tunneling parameters, to construct a geological-shield-subgrade interaction model, and to generate structural response feature data. The prediction module is used to predict the evolution based on the structural response characteristic data, the traffic load data and the historical settlement data, and generate a multi-stage deformation prediction field. The analysis module is used to perform risk analysis based on the deformation prediction field and the meteorological and hydrological data, and calculate the joint probability distribution. The generation module generates a risk control strategy sequence based on the joint probability distribution and real-time monitoring data; The risk analysis based on the deformation prediction field and the meteorological and hydrological data includes: Based on the deformation prediction field and meteorological and hydrological data, seepage-deformation coupling is performed, and a multi-field coupled deformation field is generated by quantifying the enhancing effect of pore water pressure on soil softening. Risk factors are correlated based on the multi-field coupled deformation field and the three-dimensional geological parameters, and a conditional probability relationship diagram is generated by constructing a Bayesian network topology. Risk propagation calculations are performed based on the conditional probability relationship diagram. By integrating the seepage mutation mechanism triggered by rainfall intensity threshold and the cumulative damage effect associated with traffic load frequency, a spatiotemporal joint probability distribution matrix of roadbed instability, segment cracking and surface collapse is generated. The risk factor correlation based on the multi-field coupled deformation field and the three-dimensional geological parameters includes: Risk indicators are extracted based on the multi-field coupled deformation field and the three-dimensional geological parameters to obtain a risk feature vector reflecting the structural damage sensitivity. Based on the risk feature vector, causal topology modeling is performed, and an initial directed acyclic graph containing the causal relationship between excessive settlement and segment cracking is generated by utilizing the physical correlation rules between the deterioration of the subgrade fill modulus and the stress concentration of the tunnel segments. Dynamic probability learning is performed based on the initial directed acyclic graph. By quantifying the conditional dependence strength of risk events within different geological units, a Bayesian network topology graph is output.

7. The safety risk control system for shield tunnels passing under highway subgrades according to claim 6, characterized in that, The modeling module includes: The first modeling unit is used to model the geological-subgrade interface based on the three-dimensional geological parameters and the subgrade layering characteristics to obtain the coupled stress field. The second modeling unit is used to perform shield-soil contact modeling based on the shield tunneling parameters and the coupled stress field, and to generate a dynamic contact stress field by fusing the dynamic parameters of the cutterhead torque with the characteristics of the propulsion speed gradient. The third modeling unit is used to perform multi-field coupling based on the coupled stress field and the dynamic contact stress field, and obtain structural response characteristic data by jointly calculating the seepage pressure gradient, soil stress redistribution and segment deformation coordination relationship.

8. The safety risk control system for shield tunnels passing under highway subgrades according to claim 6, characterized in that, The prediction module includes: The first prediction unit is used to decouple the dynamic response based on the structural response characteristic data and the traffic load data, and generate a dynamic stress-strain evolution sequence by separating the transient vibration and steady-state creep components of the soil. The second prediction unit is used to perform parameter inversion based on the dynamic stress-strain evolution sequence and the historical settlement data to obtain the time-varying constitutive equation of the material. The third prediction unit is used to make predictions based on the time-varying constitutive equation of the material and the dynamic stress-strain evolution sequence. By simulating the synergistic effect of soil pore water pressure dissipation and structural stiffness degradation under traffic load cycles, a multi-stage deformation prediction field is generated.

Images