Multi-dimensional comprehensive prediction and characterization method and system for tunnel gushing water in water-rich sandy gravel stratum

Abstract

The invention discloses a multi-dimensional comprehensive prediction and characterization method and system for water-rich sandy gravel stratum tunnel water gushing, and relates to the technical field of tunnel engineering.The method comprises the steps that geological survey is conducted on a water-rich sandy gravel stratum tunnel construction area, geological data are obtained, multiple monitoring sensors are arranged, and hydrodynamics and hydrochemistry data are collected in real time; building physical models of the tunnel area, wherein the physical models comprise a three-dimensional geologic model, a hydrodynamic model and a hydrochemical model; based on the hydrodynamic model and the hydrochemical model, performing dynamic water gushing prediction according to hydrodynamic data and hydrochemical data acquired in real time, and outputting predicted water gushing intensity, spatial distribution and time change trend of a next time step; and the prediction data are mapped into a three-dimensional geologic model, a flow field vector diagram and a concentration distribution heat map are generated, and a prediction result is compared with monitoring data of a next time step, so that model parameters are dynamically adjusted and updated. According to the invention, high-precision dynamic water gushing prediction can be realized.

Description

Technical Field

[0001] This invention relates to the field of tunnel engineering technology, and in particular to a multi-dimensional comprehensive prediction and characterization method and system for water inrush in tunnels in water-rich sandy and gravelly strata. Background Technology

[0002] The statements in this section are merely background information related to the present invention and do not necessarily constitute prior art.

[0003] During tunnel construction, water-rich sandy and gravelly strata are often high-risk areas for water inrush problems. Water inrush not only affects the progress and safety of tunnel construction but can also lead to geological disasters, equipment damage, and casualties among construction workers. Therefore, accurate prediction and control of water inrush is a crucial task in tunnel engineering. Traditional water inrush prediction methods can be categorized into four types: prediction methods based on geological exploration data, prediction methods based on hydrogeological modeling, single analysis methods based on hydraulic parameters, and analysis methods based on water chemical characteristics. However, water inrush is the result of the combined effects of multiple factors, including the dynamic characteristics of water flow, the source of groundwater, and its chemical composition. The aforementioned traditional methods typically rely solely on one type of data, such as geological or hydrological data, neglecting the multidimensional characteristics of water flow and the mutual influence between multiple data sources. This results in incomplete predictions and low accuracy in water inrush prediction. Furthermore, traditional methods are usually based on historical data or static analysis, making it difficult to reflect the dynamic changes of water inrush in real time and unable to cope with the rapid changes in water inrush problems under complex geological environments.

[0004] With the continuous development of tunnel construction technology, especially in complex water-rich sandy and gravelly strata, traditional prediction methods can no longer meet the demand for high-precision prediction. Therefore, more and more research is focusing on multi-dimensional data fusion techniques, combining hydrodynamic parameters, geological background information, and hydrochemical characteristics for more comprehensive analysis and prediction. However, although existing methods consider the dynamics and hydrochemical characteristics of water flow to some extent, these methods still have many shortcomings, including: (1) Existing monitoring methods rely on a small number of sensors. Monitoring points are usually set up and selected inside tunnels, which cannot actually fully cover all potential water inrush risk areas, resulting in the inability to predict and detect water inrush risks in some key areas in a timely manner. (2) Existing prediction methods often rely on fixed historical data, which cannot dynamically track changes in the water inrush area and cannot achieve real-time and accurate prediction of the water inrush process; (3) Existing prediction models are usually based on empirical formulas, which lack flexibility and cannot adapt to changes in different tunnel conditions, resulting in poor prediction accuracy. Summary of the Invention

[0005] To address the shortcomings of existing technologies, this invention provides a multi-dimensional comprehensive prediction and characterization method and system for water inrush in tunnels located in water-rich sandy gravel strata. By comprehensively considering spatial distribution and water inrush hazard, a dense sensor network is rationally deployed at multiple key locations, including the tunnel, surrounding geological layers, and potential water inrush areas. This network comprehensively considers real-time monitoring data of water flow, water pressure, and water chemical composition, combined with geological background information, and constructs a comprehensive and dynamic water inrush prediction model through multi-dimensional data fusion. This achieves high-precision and dynamic water inrush prediction, providing accurate water inrush prediction for the construction of tunnels in water-rich sandy gravel strata.

[0006] In a first aspect, the present invention provides a multi-dimensional comprehensive prediction and characterization method for water inrush in tunnels in water-rich sandy and gravelly strata.

[0007] A multi-dimensional comprehensive prediction and characterization method for water inrush in tunnels in water-rich sandy and gravelly strata includes: A geological survey was conducted in the tunnel construction area of ​​water-rich sandy and gravelly strata to obtain geological data of the tunnel area, and various monitoring sensors were deployed to collect hydrodynamic and hydrochemical data in real time. Construct a physical model of the tunnel area, including a three-dimensional geological model, a hydrodynamic model, and a hydrochemical model; Based on hydrodynamic and hydrochemical models, dynamic inrush prediction is performed using real-time collected hydrodynamic and hydrochemical data, and the predicted inrush intensity, spatial distribution, and temporal variation trend for the next time step are output. The predicted inflow intensity and spatial distribution of the next time step are mapped onto the three-dimensional geological model to generate a flow field vector map and a concentration distribution heat map. The prediction results are then compared with the monitoring data of the next time step to dynamically adjust and update the model parameters.

[0008] A further technical solution is provided, wherein the geological data includes: underground stratigraphic structure, hydrogeological conditions, and geological characteristic data; wherein, the underground stratigraphic structure includes the thickness and permeability coefficient of the sand and gravel layer; the hydrogeological conditions include groundwater level, hydraulic gradient, and aquifer distribution; and the geological characteristics include fault zones and potential karst water inflow channels. The hydrodynamic data includes: water level, water flow rate, flow velocity, and water pressure parameters; the hydrochemical data includes: ion concentration, pH value, redox potential, and dissolved solids parameters in the water body.

[0009] A further technical solution, the construction of the three-dimensional geological model, includes: A three-dimensional geological model of the tunnel area was established using three-dimensional geological modeling tools. By integrating the acquired geological data as input, a three-dimensional geological model is generated, which includes the distribution of sand and gravel strata, fault zones, aquifers, and karst formations.

[0010] A further technical solution, the construction of the hydrodynamic model and the hydrochemical model, includes: Based on the principles of seepage mechanics, a mathematical model describing groundwater flow is established. Combined with fluid mechanics equations, a set of partial differential equations describing water flow is established as the mathematical foundation of the model. Among them, the fluid mechanics equations include Darcy's law and the continuity equation. Based on the convection-diffusion-reaction equation, the mathematical foundation equations for the water chemistry model are established to simulate the migration and reaction of chemical substances in groundwater. The data-based equations of the coupled hydrodynamic and hydrochemical models are used to construct a coupled model for analyzing the effects of chemical reactions on permeability and the feedback of flow rate on changes in hydrochemistry. The coupled model was solved using the finite element method to obtain the hydrodynamic equations and hydrochemical equations, which were then used to construct the hydrodynamic and hydrochemical models as dynamic inrush prediction models.

[0011] A further technical solution involves using the finite element method to solve the coupled model, yielding the hydrodynamic equations and hydrochemical equations, including: Using the coupled model equations as the governing equations, the tunnel region is defined based on the three-dimensional geological structure model, and the boundary conditions and initial conditions of the region are defined. The tunnel region is discretized, and the governing equations are solved using the weighted residual method to obtain the weak form, namely the hydrodynamic equations and the hydrochemical equations.

[0012] A further technical solution, the dynamic inrush prediction, is as follows: The hydrodynamic and hydrochemical equations are discretized to form discrete equations, and the stiffness matrix K of the hydrodynamic and hydrochemical equations of the domain unit is calculated. e and load vector F e ; The stiffness matrix K of all elements e and load vector F e The data are assembled into a global matrix K and a load vector F. Combined with real-time acquired hydrodynamic and hydrochemical data, time discretization and solution are performed to obtain the prediction equations for water head in hydrodynamics and chemical component concentration in hydrochemistry. Based on the prediction equation, for each time step, the global linear equation system is solved to predict and update the head and chemical component concentrations for the next time step.

[0013] Secondly, the present invention provides a multi-dimensional comprehensive prediction and characterization system for water inrush in tunnels in water-rich sandy and gravelly strata.

[0014] A multi-dimensional comprehensive prediction and characterization system for water inrush in tunnels in water-rich sandy and gravelly strata includes: The data acquisition module is used to conduct geological surveys of the tunnel construction area in water-rich sandy and gravelly strata, acquire geological data of the tunnel area, and deploy various monitoring sensors to collect hydrodynamic and hydrochemical data in real time. The physical modeling module is used to build physical models of the tunnel area, including three-dimensional geological models, hydrodynamic models, and hydrochemical models. The dynamic inrush prediction module is used to make dynamic inrush predictions based on hydrodynamic and hydrochemical models and real-time collected hydrodynamic and hydrochemical data, and output the predicted inrush intensity, spatial distribution and temporal variation trend for the next time step. The water inrush characterization module is used to map the predicted water inrush intensity and spatial distribution of the next time step into the three-dimensional geological model, generate flow field vector map and concentration distribution heat map, and compare the prediction results with the monitoring data of the next time step to dynamically adjust and update the model parameters.

[0015] Thirdly, the present invention also provides an electronic device, comprising: a memory for storing executable instructions; and a processor for executing the executable instructions stored in the memory to realize the above-mentioned method for multi-dimensional comprehensive prediction and characterization of water inrush in tunnels in water-rich sandy gravel strata.

[0016] Fourthly, the present invention also provides a computer-readable storage medium storing executable instructions, which, when executed by a processor, implement the above-mentioned method for multi-dimensional comprehensive prediction and characterization of water inrush in tunnels in water-rich sandy gravel strata.

[0017] Fifthly, the present invention also provides a computer program product, which includes executable instructions stored in a computer-readable storage medium; wherein, when the processor of the electronic device reads the executable instructions from the computer-readable storage medium and executes the executable instructions, the above-mentioned method for multi-dimensional comprehensive prediction and characterization of water inrush in tunnels in water-rich sandy gravel strata is realized.

[0018] The above one or more technical solutions have the following beneficial effects: 1. This invention provides a multi-dimensional comprehensive prediction and characterization method and system for water inrush in tunnels in water-rich sandy gravel strata. By comprehensively considering spatial distribution and water inrush hazard, a dense sensor network is rationally deployed at multiple key locations such as the tunnel and surrounding geological layers and potential water inrush areas. This network comprehensively considers real-time monitoring data of water flow, water pressure, and water chemical composition, combined with geological background information, and constructs a comprehensive and dynamic water inrush prediction model through multi-dimensional data fusion. This model not only considers the physical characteristics of water flow but also incorporates the chemical characteristics of water, enabling more comprehensive and high-precision dynamic water inrush prediction. It comprehensively and accurately predicts the intensity, time, and spatial distribution of water inrush, significantly improving the accuracy and reliability of prediction, and providing accurate water inrush prediction for the construction of tunnels in water-rich sandy gravel strata.

[0019] 2. Compared to traditional monitoring and prediction methods that rely on only a few single sensors, this application deploys a dense sensor network in the tunnel construction area and surrounding potential water inrush risk areas. That is, the deployment is not limited to the tunnel itself, but also includes multiple key locations such as the geological layers around the tunnel, potential water inrush areas, groundwater levels, and groundwater flow directions. By comprehensively considering spatial distribution and water inrush risk, sensors are rationally arranged, and multiple sensors work together to ensure real-time monitoring of all potential water inrush sources and their paths. This dense deployment of sensors helps to fill the monitoring blind spots in traditional methods and achieve comprehensive coverage of water inrush risk areas.

[0020] 3. This invention dynamically optimizes and adjusts the prediction model by combining real-time monitoring data. This allows the prediction results to be updated in real time according to the actual situation on site and changes in water inflow, and optimizes risk assessment and intervention measures, making prediction and risk management more flexible, real-time and accurate.

[0021] 4. This invention employs water chemical fingerprinting technology. By analyzing the chemical components such as ion concentration, pH value, and isotopes in water samples, it can accurately identify the specific source of water inflow, such as groundwater, surface water seepage, and karst water. This method not only improves the accuracy of water source identification but also provides more accurate input data for subsequent water inflow path prediction.

[0022] 5. This invention combines hydrodynamic and hydrochemical models to accurately predict water inflow paths and performs spatial visualization based on a three-dimensional geological model. It can not only identify the direction of water inflow channels, but also draw a spatial distribution map of water inflow in real time, intuitively displaying high-risk areas and providing tunnel construction personnel with a scientific basis for risk assessment and construction decision-making.

[0023] Advantages of additional aspects of the invention will be set forth in part in the description which follows, and in part will be obvious from the description, or may be learned by practice of the invention. Attached Figure Description

[0024] The accompanying drawings, which form part of this invention, are used to provide a further understanding of the invention. The illustrative embodiments of the invention and their descriptions are used to explain the invention and do not constitute an improper limitation of the invention.

[0025] Figure 1 This is an overall flowchart of the multi-dimensional comprehensive prediction and characterization method for water inrush in tunnels in water-rich sandy and gravelly strata as described in this embodiment of the invention. Detailed Implementation

[0026] It should be noted that the following detailed descriptions are exemplary and are intended only to describe specific embodiments and to provide further explanation of the invention, and are not intended to limit the scope of exemplary embodiments of the invention. Unless otherwise specified, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains. Furthermore, it should be understood that when the terms "comprising" and / or "including" are used in this specification, they indicate the presence of features, steps, operations, devices, components, and / or combinations thereof.

[0027] Example 1 This embodiment provides a multi-dimensional comprehensive prediction and characterization method for water inrush in tunnels located in water-rich sandy and gravelly strata, such as... Figure 1 As shown, the specific steps include: Step S1: Conduct a geological survey of the tunnel construction area in the water-rich sandy and gravelly strata, obtain geological data of the tunnel area, and deploy various monitoring sensors to collect hydrodynamic and hydrochemical data in real time.

[0028] Specifically, a detailed underground geological survey is conducted in the tunnel construction area of ​​water-rich sandy and gravelly strata. This includes obtaining geological data of the area through geological exploration (such as drilling, geological profiles, etc.), resistivity detection, and seismic wave detection. This geological data includes data on underground strata structure (thickness of sandy and gravelly layers, permeability coefficient, etc.), hydrogeological conditions (such as groundwater level, hydraulic gradient, aquifer distribution, etc.), and geological features (such as fault zones, karst, and other potential water inflow channels).

[0029] Furthermore, based on the geological survey results, various hydrodynamic sensors, such as water level sensors, flow meters, velocity meters, and pressure sensors, are deployed in the tunnel construction area to monitor hydrodynamic parameters such as water level, flow rate, velocity, and pressure in real time. Similarly, various water chemical monitoring devices are deployed, and chemical sensors such as ion concentration meters and pH meters are installed to collect and acquire water chemical parameters such as ion concentration, pH value, redox potential, and dissolved solids in the water in real time.

[0030] Preferably, a real-time data transmission system is established to transmit monitoring data to a remote central processing unit for subsequent data analysis and processing. Various sensors measure the physical changes (such as flow velocity and pressure) and chemical composition (such as ion concentration) of the water flow in real time. The sampling frequency of each sensor can be set according to specific needs, such as once per minute or once per second. These sensors continuously collect hydrodynamic and hydrochemical parameter data within the tunnel area and send the data to a sensor network gateway. The sensor network gateway receives and initially processes the raw data, performing noise filtering and format conversion, before transmitting it to the central processing unit via wired or wireless network. A reliable data transmission protocol (such as HTTP) can be used during transmission to ensure the accuracy and real-time performance of the data.

[0031] Step S2: Construct a physical model of the tunnel area, including a three-dimensional geological model, a hydrodynamic model, and a hydrochemical model.

[0032] Step S2.1: First, use a 3D geological modeling tool (such as RockWorks) to establish a 3D geological model of the tunnel area. Integrate the geological, drilling, and geophysical data obtained in the previous step as the input layer to generate a 3D geological model of the sand and gravel strata, fault zone, aquifer, and karst distribution, and output it as the output layer.

[0033] Step S2.2: Next, perform hydrodynamic and hydrochemical modeling.

[0034] Step S2.2.1: Based on the principles of seepage mechanics and fluid mechanics equations, establish the mathematical foundation equations of the hydrodynamic model to describe groundwater flow. Specifically, based on the principles of seepage mechanics, establish a mathematical model describing groundwater flow, and combine it with the fluid mechanics equations to establish a set of partial differential equations describing water flow, serving as the mathematical basis of the model. The fluid mechanics equations include Darcy's law and the continuity equation.

[0035] Darcy's Law is as follows: ; In the above formula, q For seepage flow rate, k Permeability coefficient, h This refers to the water head height.

[0036] The continuity equation is: ; In the above formula, The water storage coefficient, This represents the rate of change of water head over time.

[0037] Based on this, the fundamental partial differential equation describing three-dimensional groundwater flow is established as follows: ; In the above formula, Let be the water head height, which is an unknown quantity; k x 、k y 、k z These are the permeability coefficients in different directions (anisotropy is considered here); S s The water storage coefficient describes the amount of water stored per unit volume of aquifer due to changes in hydraulic head. W These are source and sink terms, representing the rate at which external water flows into (e.g., through surface seepage) or out of (e.g., through pumping) a system. t It is a time variable.

[0038] Step S2.2.2: Based on the convection-diffusion-reaction equation, establish the mathematical foundation equations for the water chemistry model to simulate the migration and reaction of chemical substances in groundwater. Specifically, the governing equations for chemical substance migration and reaction are shown below, which describe the movement of chemical components in the aquifer due to fluid convection (…). ),diffusion( ) and reaction ( The effect of ) is expressed in the following formula: ; In the above formula, It is the concentration of a chemical component (such as ion concentration), which is an unknown quantity; , representing the velocity vector of groundwater (provided by the hydrodynamic model); K To affect the permeability coefficient; D It is the diffusion coefficient of chemical substances (which takes into account porosity and formation characteristics); This is a chemical reaction term, representing the contribution of the concentration change to the chemical reaction.

[0039] Step S2.2.3: Couple the hydrodynamic model and the hydrochemical model with their underlying data equations to construct a coupled model for analyzing the effects of chemical reactions on permeability and the feedback of flow rate on changes in hydrochemistry.

[0040] Specifically, in predicting tunnel water inrush, the hydrodynamic model and the hydrochemical model need to be solved simultaneously. Therefore, a coupled model is established by simultaneously solving the above equations. The main equations of this coupled model are Darcy's law, the basic partial differential equations, and the convection-diffusion-reaction equations. There are two coupling points: one is that the water flow velocity affects chemical migration; the other is that the flow velocity... q Provided by the hydrodynamic model, it directly affects the convection term in the hydrochemical model. Secondly, chemical changes feed back into hydraulic characteristics. Chemical reactions may cause precipitation or dissolution, such as carbonate precipitation. This characteristic affects the permeability coefficient. K) and diffusion coefficient ( D ).

[0041] Step S2.2.4: Solve the coupled model using the finite element method (FEM) to obtain the hydrodynamic equation and the hydrochemical equation, and use these to construct the hydrodynamic model and the hydrochemical model as a dynamic inrush prediction model.

[0042] Specifically, firstly, the problem is described and equations are established, including: determining the governing equations (i.e., the aforementioned coupled equation set), defining the region (i.e., the three-dimensional geological structure model Ω) and boundary conditions, and initial conditions. The boundary conditions include: for hydrodynamics, given the boundary water pressure (…). h ) or permeation boundary flux ( For water chemistry, given a boundary concentration ( C ) or chemical flux ( Initial conditions include: initial head distribution. Initial concentration distribution .

[0043] Then, finite element discretization is performed, and the tunnel region Ω is discretized (i.e., meshed into multiple element regions). During the discretization process, a piecewise linear function is used: As a trial function, where For shape functions, i For the first i Each unit area.

[0044] Then, the weighted residual method was used to solve the governing equations, yielding the weak-form hydrodynamic and hydrochemical equations, which are expressed as follows: ; ; in, Let be the weighting function, and calculate the above integral term within the cell.

[0045] Step S3: Based on the hydrodynamic model and the hydrochemical model, perform dynamic inrush prediction according to the real-time collected hydrodynamic data and hydrochemical data, and output the predicted inrush intensity, spatial distribution and temporal variation trend for the next time step.

[0046] Specifically, the hydrodynamic equations and hydrochemical equations are discretized to form discrete equations, and the stiffness matrix K of the calculation unit is then used. e and load vector F e Among them, the stiffness matrix K of hydrodynamics e With load vector F e for: ; ; Stiffness matrix K of water chemistry e With load vector F e for: ; .

[0047] Then, the stiffness matrix K of all elements is... e and load vector F e Assembled into a global matrix K and a load vector F, according to , By combining real-time acquired hydrodynamic and hydrochemical data, nodes with known head or concentration are directly assigned values ​​and their corresponding matrix rows and columns are adjusted. Here, H and C are the global vectors for head and concentration, respectively; K... h and K c K is the stiffness matrix, which corresponds to the system stiffness of the hydrodynamic (h) and hydrochemical (C) models, respectively. h K is the hydrodynamic stiffness matrix, which describes the spatial variation of water flow (head) distribution and is related to factors such as water flow gradient, permeability coefficient, and boundary conditions. c The water chemical stiffness matrix describes the spatial distribution of chemical components (such as ion concentrations). This parameter takes into account factors such as chemical reactions, diffusion, and convection; F h and F c F represents the load vector, i.e., the effect of external forces (or source terms) at each node. h This is the hydrodynamic load vector, representing the source term of external water flow influence, such as external water sources, seepage flow, etc.; F c This is the water chemical load vector, which represents external chemical source terms, such as external dissolved substances, reaction sources, etc.

[0048] Next, time discretization and solution are performed, that is, the backward Euler method is used to discretize the time term to obtain the prediction equations for water head in hydrodynamics and chemical component concentration in hydrochemistry, as follows: ; .

[0049] Based on the above equations, solve the global linear equation system K for each time step. h H n+1 =F h and K C C n+1 =F C Predict the head H at the next time step n+1 and chemical component concentration C n+1 And update, among which n Indicates the current moment or the current time step.n +1 indicates the next moment or the next time step.

[0050] As another implementation, this embodiment can also identify the source of water inrush. The analysis of identifying the source of water inrush is an optimization step for the water inrush prediction results and is one of the important links in the prediction model optimization and parameter update.

[0051] In this embodiment, identifying the source of the water inrush includes hydrochemical fingerprinting, hydrodynamic gradient analysis, and mixing analysis, wherein: (1) Aquatic chemical fingerprinting: This method identifies the type of water source by analyzing ion concentrations, pH values, and isotopes in water samples. Specifically, it uses aquatic chemical data, primarily ion concentrations (such as Na+). + Ca 2+ HCO3 - Cl - The water body's pH value, redox potential, water temperature, and dissolved gas concentration (such as CO2, O2, etc.) are analyzed to determine its chemical characteristics (such as ionic composition characteristics). The water body is compared with known water samples (such as surface water, groundwater, karst water, etc.) to identify the source of the water body and its connection with different geological units.

[0052] (2) Hydrodynamic gradient analysis: By analyzing the water level gradient and velocity field, the direction and source of water flow are determined. Specifically, by collecting hydrodynamic data (such as water pressure, velocity, and gradient), the flow characteristics and gradient distribution of groundwater around the tunnel are obtained, and the driving force of water inrush is determined.

[0053] (3) Mixed analysis: Combining multi-source water chemistry data, the contribution ratio of each water source is estimated using the mass balance method, where the mass balance formula is: ,in, For source type i The proportion of contribution, For source type i The chemical concentration.

[0054] In addition, geological background data, such as tunnel geological profiles, aquifer distribution, and the location of faults and karst channels, are used to provide boundary conditions for fluid flow; water level contour lines in the area surrounding the tunnel are calculated to analyze the direction of water flow and the possible location of water sources; and groundwater flow models (i.e., partial differential equations) are used to predict the connection between water inrush points and water source areas.

[0055] By employing the aforementioned techniques such as hydrochemical fingerprinting, hydrodynamic gradient analysis, and hybrid analysis, the source of water inrush (such as groundwater, surface water, karst water, etc.) can be accurately determined. Based on the identification results, the intensity and flow path of the water inrush can be optimized and adjusted. Furthermore, the prediction model can be optimized and its parameters updated based on the optimized and adjusted data, thereby improving the accuracy of water inrush prediction and providing more refined input data for subsequent model updates.

[0056] Furthermore, based on the different nature and source of the water inrush, targeted prevention and control measures can be taken to improve the scientific nature and effectiveness of water inrush risk management.

[0057] Furthermore, in this embodiment, considering that predicting water inrush solely based on the constructed hydrodynamic and hydrochemical models is achieved through physical equations (such as Darcy's law, continuity equations, etc.) and hydrogeological data, this physical model has certain parameter dependencies and lacks flexibility, resulting in suboptimal prediction accuracy. The parameter dependencies include: many parameters in the physical model, such as permeability coefficients and hydraulic head distribution, may have significant uncertainties, and the measurement and estimation of these parameters typically require substantial field data or assumptions; the lack of flexibility is also due to the fact that physical models generally operate under assumed geological conditions and flow patterns, making it difficult to handle particularly complex and dynamically changing actual engineering environments. Based on the above, this embodiment introduces machine learning and statistical modeling techniques to overcome the limitations of the physical model, further improving the accuracy and dynamic adaptability of the prediction through a data-driven approach.

[0058] Therefore, preferably, while using the constructed hydrodynamic and hydrochemical models (i.e., physical models) to predict water inrush as described above, this embodiment also constructs a dynamic water inrush prediction model, that is, using machine learning or statistical modeling techniques to construct a dynamic water inrush prediction model. Specifically, in this model, real-time collected hydrodynamic, hydrochemical, and geological information data are input, and the predicted water inrush intensity, spatial distribution, and temporal variation trend are output. The model is validated and optimized using historical data, and the model parameters are dynamically adjusted based on actual monitoring results to improve prediction accuracy.

[0059] In this process, the physical model and the machine learning model work collaboratively. The physical model provides the initial framework for prediction, describing the basic physical mechanisms and processes of water inrush, while the machine learning model optimizes the parameters of the physical model and corrects the water inrush prediction results based on historical and real-time data. Specifically, based on the machine learning model (i.e., the dynamic water inrush prediction model), and using real-time hydrodynamic, hydrochemical, and geological information data, the predicted water inrush intensity, spatial distribution, and temporal variation trend for the next time step are predicted. The prediction results output by this model are then used to correct the water inrush prediction results generated by the physical models (i.e., the hydrodynamic model and the hydrochemical model) to obtain more accurate water inrush prediction results. This correction can be achieved using methods such as weighted averaging, which will not be described in detail here.

[0060] By introducing machine learning models, the accuracy and adaptability of predictions can be improved: (1) Model nonlinearity: Machine learning can automatically learn complex nonlinear relationships from data by processing a large amount of historical data. This is especially important for complex geological environments and dynamically changing water inrush phenomena. Machine learning models can identify details that physical models fail to capture, such as the nonlinear relationship between water flow and hydrochemical characteristics under specific conditions; (2) Dynamic adaptability: Machine learning models have adaptive capabilities and can dynamically adjust model parameters based on real-time collected hydrodynamic data, hydrochemical data and geological data to improve prediction accuracy. For example, with the input of new real-time data, machine learning models can promptly correct the predicted value and spatial distribution of water inrush intensity.

[0061] By introducing machine learning models, physical models can be optimized and verified: (1) Historical data verification: Although physical models provide theoretical prediction results, their accuracy is often affected by factors such as field conditions and parameter estimation. Therefore, by using historical data as input, machine learning models can verify and optimize physical models based on actual monitoring data, make up for errors in physical models, and machine learning models can self-adjust within a certain range to adapt to new data patterns and changes; (2) Self-adjustment and optimization: Machine learning models can adjust weights and optimize model structure in real time during operation, so as to minimize the difference between the output of physical models and actual observation data, thereby enhancing the predictive ability of the model.

[0062] By introducing machine learning models, data fusion and complex pattern recognition can be achieved: (1) Data fusion: Machine learning can process and fuse data from multiple sources (such as hydrodynamic, hydrochemical, geological data, etc.), extract the most meaningful features, and further refine the water inrush prediction; (2) Complex pattern recognition: Through a large amount of historical data and real-time monitoring data, machine learning algorithms can identify the changes in water inrush intensity, spatial distribution and complex patterns between water inrush and environmental factors, which are difficult to capture by traditional physical models.

[0063] By combining the aforementioned physical model with a data-driven model, we can better address uncertainties in complex geological environments, thereby obtaining more accurate and flexible water inrush predictions.

[0064] Step S4: Map the predicted inflow intensity and spatial distribution of the next time step into the three-dimensional geological model to generate a flow field vector map and a concentration distribution heat map. Compare the prediction results with the monitoring data of the next time step to dynamically adjust and update the model parameters.

[0065] Specifically, based on the prediction results, the formation causes, evolution process and possible expansion trends of the water inrush channel are analyzed, and a three-dimensional spatial characterization is performed.

[0066] First, using known hydrodynamic data and the output of the hydrodynamic model, the velocity and direction fields are plotted, and then the flow path is drawn to identify potential inrush channels. Based on the analysis of changes in known hydrochemical data, the flow path and characteristics are inferred, such as high... It may flow through karst areas; high It may pass through a sulfur-containing mineral layer; high , The source of the water inflow could be surface water seepage or deep water upwelling. This allows for reverse source tracing analysis, which involves inferring the path and residence time of the water flow based on its velocity and direction. This includes identifying potential water inflow paths using geological models, such as rock fissures and faults, karst channels, and highly permeable sand and gravel layers.

[0067] Secondly, the obtained water head and concentration distribution results are mapped back to the 3D geological model to generate flow field vector maps, concentration distribution heatmaps, etc. These numerical results are then compared with monitoring data to adjust model parameters, such as permeability coefficient and diffusion coefficient. Finally, combined with updated monitoring data, the model's predictions of the spatiotemporal distribution of tunnel water inflow intensity and chemical characteristics are dynamically adjusted. Additionally, based on the predicted water inflow intensity and distribution, the water inflow risk level in different areas is assessed, and a risk assessment report is generated.

[0068] The spatial distribution of water flow paths and water inflow channels is visualized in a three-dimensional geological model using the above method, helping engineers to intuitively understand the water inflow area.

[0069] Example 2 This embodiment provides a multi-dimensional comprehensive prediction and characterization system for water inrush in tunnels in water-rich sandy and gravelly strata, including: The data acquisition module is used to conduct geological surveys of the tunnel construction area in water-rich sandy and gravelly strata, acquire geological data of the tunnel area, and deploy various monitoring sensors to collect hydrodynamic and hydrochemical data in real time. The physical modeling module is used to build physical models of the tunnel area, including three-dimensional geological models, hydrodynamic models, and hydrochemical models. The dynamic inrush prediction module is used to make dynamic inrush predictions based on hydrodynamic and hydrochemical models and real-time collected hydrodynamic and hydrochemical data, and output the predicted inrush intensity, spatial distribution and temporal variation trend for the next time step. The water inrush characterization module is used to map the predicted water inrush intensity and spatial distribution of the next time step into the three-dimensional geological model, generate flow field vector map and concentration distribution heat map, and compare the prediction results with the monitoring data of the next time step to dynamically adjust and update the model parameters.

[0070] Example 3 This embodiment provides an electronic device, including: a memory for storing executable instructions; and a processor for executing the executable instructions stored in the memory to implement the method provided in this embodiment.

[0071] Example 4 This embodiment also provides a computer-readable storage medium storing executable instructions, which, when executed by a processor, will cause the processor to execute the method described above in this embodiment.

[0072] Example 5 This embodiment provides a computer program product including executable instructions, which are computer instructions; the executable instructions are stored in a computer-readable storage medium. When the processor of an electronic device reads the executable instructions from the computer-readable storage medium and executes the executable instructions, the electronic device performs the method described in this embodiment.

[0073] The steps and methods involved in Embodiments 2 to 5 above correspond to those in Embodiment 1. For specific implementation details, please refer to the relevant description section of Embodiment 1. The term "computer-readable storage medium" should be understood as a single medium or multiple media including one or more instruction sets; it should also be understood as including any medium capable of storing, encoding, or carrying an instruction set for execution by a processor and enabling the processor to perform any of the methods in this invention.

[0074] Those skilled in the art will understand that the modules or steps of the present invention described above can be implemented using general-purpose computer devices. Optionally, they can be implemented using computer-executable program code, thereby allowing them to be stored in a storage device for execution by a computer device, or they can be fabricated as separate integrated circuit modules, or multiple modules or steps can be fabricated as a single integrated circuit module. The present invention is not limited to any particular combination of hardware and software.

[0075] The above description is only a preferred embodiment of the present invention. Although the specific implementation of the present invention has been described in conjunction with the accompanying drawings, it is not intended to limit the scope of protection of the present invention. Those skilled in the art should understand that, based on the technical solution of the present invention, various modifications or variations that can be made by those skilled in the art without creative effort are still within the scope of protection of the present invention.

Claims

1. A multi-dimensional comprehensive prediction and characterization method for water inrush in tunnels in water-rich sandy and gravelly strata, characterized in that, include: A geological survey was conducted in the tunnel construction area of ​​water-rich sandy and gravelly strata to obtain geological data of the tunnel area, and various monitoring sensors were deployed to collect hydrodynamic and hydrochemical data in real time. Construct a physical model of the tunnel area, including a three-dimensional geological model, a hydrodynamic model, and a hydrochemical model; Based on hydrodynamic and hydrochemical models, dynamic inrush prediction is performed using real-time collected hydrodynamic and hydrochemical data, and the predicted inrush intensity, spatial distribution, and temporal variation trend for the next time step are output. The predicted inflow intensity and spatial distribution of the next time step are mapped onto the three-dimensional geological model to generate a flow field vector map and a concentration distribution heat map. The prediction results are then compared with the monitoring data of the next time step to dynamically adjust and update the model parameters.

2. The method for multi-dimensional comprehensive prediction and characterization of water inrush in tunnels in water-rich sandy and gravelly strata as described in claim 1, characterized in that, The geological data includes: underground stratigraphic structure, hydrogeological conditions, and geological feature data; among which, underground stratigraphic structure includes the thickness and permeability coefficient of sand and gravel layers; hydrogeological conditions include groundwater level, hydraulic gradient, and aquifer distribution; and geological features include fault zones and potential karst water inflow channels. The hydrodynamic data includes: water level, water flow rate, flow velocity, and water pressure parameters; the hydrochemical data includes: ion concentration, pH value, redox potential, and dissolved solids parameters in the water body.

3. The method for multi-dimensional comprehensive prediction and characterization of water inrush in tunnels in water-rich sandy and gravelly strata as described in claim 1, characterized in that, The construction of the three-dimensional geological model includes: A three-dimensional geological model of the tunnel area was established using three-dimensional geological modeling tools. By integrating the acquired geological data as input, a three-dimensional geological model is generated, which includes the distribution of sand and gravel strata, fault zones, aquifers, and karst formations.

4. The method for multi-dimensional comprehensive prediction and characterization of water inrush in tunnels in water-rich sandy and gravelly strata as described in claim 1, characterized in that, The construction of the hydrodynamic and hydrochemical models includes: Based on the principles of seepage mechanics, a mathematical model describing groundwater flow is established. Combined with fluid mechanics equations, a set of partial differential equations describing water flow is established as the mathematical foundation of the model. Among them, the fluid mechanics equations include Darcy's law and the continuity equation. Based on the convection-diffusion-reaction equation, the mathematical foundation equations for the water chemistry model are established to simulate the migration and reaction of chemical substances in groundwater. The data-based equations of the coupled hydrodynamic model and the hydrochemical model are used to construct a coupled model for analyzing the effect of chemical reactions on permeability and the feedback of flow rate on changes in hydrochemistry. The coupled model was solved using the finite element method to obtain the hydrodynamic equations and hydrochemical equations, which were then used to construct the hydrodynamic and hydrochemical models as dynamic inrush prediction models.

5. The method for multi-dimensional comprehensive prediction and characterization of water inrush in tunnels in water-rich sandy and gravelly strata as described in claim 4, characterized in that, The coupled model was solved using the finite element method, yielding the hydrodynamic equations and hydrochemical equations, including: Using the coupled model equations as the governing equations, the tunnel region is defined based on the three-dimensional geological structure model, and the boundary conditions and initial conditions of the region are defined. The tunnel region is discretized, and the governing equations are solved using the weighted residual method to obtain the weak form, namely the hydrodynamic equations and the hydrochemical equations.

6. The method for multi-dimensional comprehensive prediction and characterization of water inrush in tunnels in water-rich sandy and gravelly strata as described in claim 5, characterized in that, The dynamic water inrush prediction is as follows: The hydrodynamic and hydrochemical equations are discretized to form discrete equations, and the stiffness matrix K of the hydrodynamic and hydrochemical equations of the domain unit is calculated. e and load vector F e ; The stiffness matrix K of all elements e and load vector F e The data are assembled into a global matrix K and a load vector F. Combined with real-time acquired hydrodynamic and hydrochemical data, time discretization and solution are performed to obtain the prediction equations for water head in hydrodynamics and chemical component concentration in hydrochemistry. Based on the prediction equation, for each time step, the global linear equation system is solved to predict and update the head and chemical component concentrations for the next time step.

7. A multi-dimensional comprehensive prediction and characterization system for water inrush in tunnels in water-rich sandy and gravelly strata, characterized in that, include: The data acquisition module is used to conduct geological surveys of the tunnel construction area in water-rich sandy and gravelly strata, acquire geological data of the tunnel area, and deploy various monitoring sensors to collect hydrodynamic and hydrochemical data in real time. The physical modeling module is used to build physical models of the tunnel area, including three-dimensional geological models, hydrodynamic models, and hydrochemical models. The dynamic inrush prediction module is used to make dynamic inrush predictions based on hydrodynamic and hydrochemical models and real-time collected hydrodynamic and hydrochemical data, and output the predicted inrush intensity, spatial distribution and temporal variation trend for the next time step. The water inrush characterization module is used to map the predicted water inrush intensity and spatial distribution of the next time step into the three-dimensional geological model, generate flow field vector map and concentration distribution heat map, and compare the prediction results with the monitoring data of the next time step to dynamically adjust and update the model parameters.

8. An electronic device, characterized in that, include: Memory, used to store executable instructions; The processor, when executing the executable instructions stored in the memory, implements the multi-dimensional comprehensive prediction and characterization method for water inrush in tunnels in water-rich sandy gravel strata as described in any one of claims 1-6.

9. A computer-readable storage medium, characterized in that, The device stores executable instructions that, when executed by a processor, implement the multidimensional comprehensive prediction and characterization method for water inrush in tunnels in water-rich sandy gravel strata as described in any one of claims 1-6.

10. A computer program product, characterized in that, The computer program product includes executable instructions stored in a computer-readable storage medium; When the processor of the electronic device reads the executable instructions from the computer-readable storage medium and executes the executable instructions, it implements the method for multi-dimensional comprehensive prediction and characterization of water inrush in tunnels in water-rich sandy gravel strata as described in any one of claims 1-6.

Images