A method and system for judging the damage of water-bearing surrounding rock in underground engineering

Abstract

The present invention discloses a method and system for distinguishing the failure of water-bearing surrounding rock in underground engineering. The method includes: obtaining water pressure change data of water-bearing surrounding rock, calculating the stress field change trend in combination with a preset stress distribution model, and using a crack expansion model to simulate the micro-crack expansion process; judging the nonlinear characteristics of crack expansion based on the crack expansion simulation results and the physical properties of the surrounding rock; analyzing the impact of water pressure fluctuations on the permeability characteristics of the surrounding rock, and accordingly correcting the crack expansion model parameters and recalculating the dynamic process; using multi-field coupling analysis to construct a multidimensional discrimination model for surrounding rock failure, combining time evolution data to dynamically capture the critical state, and using a nonlinear dynamic algorithm to predict the development trend of failure. Through multi-dimensional data analysis and model iterative optimization, the present invention achieves accurate simulation and prediction of the surrounding rock failure process under complex geological conditions, providing important technical support for engineering safety.

Description

Technical Field

[0001] The present invention belongs to the technical field of water-bearing surrounding rock failure in underground engineering, and in particular relates to a method and system for distinguishing water-bearing surrounding rock failure in underground engineering. Background Art

[0002] In underground engineering projects, such as underground powerhouses of hydropower stations, coal mine tunnels, and subway tunnels, the stability of water-bearing surrounding rock is one of the key factors affecting engineering safety. Existing technologies generally assess the stability of surrounding rock by monitoring parameters such as water pressure, surrounding rock stress, and crack extension. These methods can, to a certain extent, reflect the stress state and failure trend of the surrounding rock. For example, water pressure monitoring can be used to understand the impact of water-bearing conditions on the surrounding rock; stress analysis can reveal the stress distribution of the surrounding rock; and crack extension models can be used to predict the direction and speed of crack development. In addition, some studies have also introduced permeability characteristics analysis to evaluate the impact of water pressure changes on the permeability of the surrounding rock. These technical means provide basic data support for surrounding rock stability assessment and can, to a certain extent, guide engineering practice.

[0003] However, existing technologies have significant deficiencies in the identification of surrounding rock failure. First, most existing methods are based on the analysis of a single physical field, ignoring the mutual coupling relationship between multiple physical fields such as water pressure, stress, crack expansion and permeability. This isolated analysis method cannot accurately reflect the complex dynamic characteristics of the surrounding rock under water conditions, resulting in limited accuracy and reliability of the identification results. Secondly, when dealing with the nonlinear characteristics of crack expansion, existing technologies often use simplified linear models, which makes it difficult to truly reflect the complex process of crack expansion. In addition, the existing methods lack the ability to capture dynamics in the time dimension and cannot effectively track the evolution of the critical state of surrounding rock failure. Summary of the Invention

[0004] In order to solve the above technical problems, the present invention proposes a method and system for determining damage to water-bearing surrounding rock in underground engineering to solve the problems existing in the above-mentioned prior art.

[0005] To achieve the above-mentioned objectives, in a first aspect, the present invention provides a method for distinguishing the damage of water-bearing surrounding rock in underground engineering, comprising: obtaining water pressure change data of water-bearing surrounding rock, and calculating the change trend of the surrounding rock stress field in combination with a preset stress distribution model; according to the change trend of the surrounding rock stress field, using a crack expansion model to simulate the expansion process and direction of microcracks in the surrounding rock; based on the simulation results of the microcrack expansion process, combined with the physical property parameters of the surrounding rock, judging the nonlinear characteristics of crack expansion; obtaining time series data of water pressure fluctuations, and analyzing the influence of water pressure on the permeability characteristics of the surrounding rock; according to the change of the permeability characteristics, correcting the parameters of the crack expansion model, and recalculating the dynamic process of crack expansion; based on the corrected dynamic process of crack expansion, using a multi-field coupling analysis method to construct a multidimensional discrimination model for surrounding rock damage; obtaining time evolution data, and combining the multidimensional discrimination model to dynamically capture the critical state of surrounding rock damage; according to the change trend of the critical state, using a nonlinear dynamic algorithm to predict the development trend of surrounding rock damage.

[0006] Preferably, the step of calculating the changing trend of the surrounding rock stress field includes: obtaining water pressure change data of the water-bearing surrounding rock and extracting the water pressure change value; extracting the stress modulus and distribution state based on the water pressure change value in combination with a preset stress distribution model; and using a pre-established calculation method to calculate the changing trend of the surrounding rock stress field.

[0007] Preferably, the steps of simulating the expansion process and direction of microcracks in the surrounding rock include: obtaining the stress field distribution state of the surrounding rock mass, extracting the stress value and its changing trend; using a crack model, combined with the stress value changing trend, to calculate the expansion direction of the microcracks; and analyzing the expansion process of the microcracks in the surrounding rock mass through the simulation process.

[0008] Preferably, the step of determining the nonlinear characteristics of crack expansion includes: generating dynamic data of crack expansion and extracting key change points through time series analysis; analyzing the nonlinear characteristic values of crack expansion in combination with the physical property parameters of the surrounding rock; determining whether the characteristic values of crack expansion reach a preset threshold and determining the stability state of the surrounding rock mass.

[0009] Preferably, the step of recalculating the dynamic process of crack expansion includes: obtaining time series data of permeability characteristics and extracting the changing trend of permeability characteristics; using a preset crack expansion model and correcting the parameter values of the model based on the changing trend of permeability characteristics; recalculating the dynamic process of crack expansion based on the corrected parameter values to obtain the evolution result of crack expansion.

[0010] Preferably, the steps of constructing a multidimensional discriminant model of surrounding rock failure by using a multi-field coupling analysis method include: obtaining corrected dynamic process data of crack expansion, and extracting characteristic indicators of surrounding rock failure by using a multi-field coupling analysis method; constructing a multidimensional discriminant model based on the characteristic indicators of surrounding rock failure, and determining the key characteristics of surrounding rock failure; and training the multidimensional discriminant model based on the key characteristics of surrounding rock failure by using a support vector machine algorithm to obtain a trained multidimensional discriminant model.

[0011] Preferably, the step of dynamically capturing the critical state of surrounding rock failure includes: obtaining time series data during the surrounding rock evolution process and extracting evolution characteristics; using a multidimensional discriminant model to dynamically capture the extracted evolution characteristics; if the dynamically captured characteristics meet the preset critical state conditions, then determining the critical state of surrounding rock failure.

[0012] Preferably, the step of predicting the development trend of surrounding rock failure includes: obtaining time series data of the surrounding rock mass and extracting characteristic values during the surrounding rock evolution process; establishing a prediction model using a nonlinear dynamic algorithm based on the changing trend of the characteristic values; and analyzing the development trend of surrounding rock failure through the prediction model.

[0013] In a second aspect, the present invention also discloses a system for distinguishing the damage of water-bearing surrounding rocks in underground engineering, the system comprising: a water pressure change data acquisition module for obtaining water pressure change data of water-bearing surrounding rocks, and calculating the change trend of the surrounding rock stress field in combination with a preset stress distribution model; a stress field calculation module for simulating the expansion process and direction of microcracks in the surrounding rock using a crack expansion model according to the change trend of the stress field; a crack expansion simulation module for judging the nonlinear characteristics of crack expansion based on the simulation results of the crack expansion process and in combination with the physical property parameters of the surrounding rock; and a nonlinear characteristic judgment module for obtaining the time series of water pressure fluctuations. The system uses the water pressure fluctuation analysis module to modify the parameters of the crack expansion model and recalculate the dynamic process of crack expansion according to the changes in the permeability characteristics. The crack expansion model modification module is used to construct a multi-dimensional discrimination model of surrounding rock failure based on the modified dynamic process of crack expansion using a multi-field coupling analysis method. The multi-dimensional discrimination model construction module is used to obtain time evolution data and dynamically capture the critical state of surrounding rock failure in combination with the multi-dimensional discrimination model. The critical state capture module is used to predict the development trend of surrounding rock failure based on the changing trend of the critical state using a nonlinear dynamic algorithm.

[0014] In a third aspect, the present invention further discloses a computer program product, comprising a computer program, which implements the steps of the method described in the first aspect when executed by a processor.

[0015] Compared with the prior art, the present invention has the following advantages and technical effects:

[0016] The present invention discloses a method for distinguishing the destruction of water-bearing surrounding rocks of underground projects. The method first obtains the water pressure change data of the water-bearing surrounding rocks, calculates the stress field change trend in combination with a preset stress distribution model, and uses a crack expansion model to simulate the micro-crack expansion process. Then, based on the crack expansion simulation results and the physical properties of the surrounding rocks, the nonlinear characteristics of the crack expansion are judged. Next, the influence of water pressure fluctuations on the permeability characteristics of the surrounding rocks is analyzed, and the crack expansion model parameters are corrected accordingly and the dynamic process is recalculated. Finally, the present invention uses multi-field coupling analysis to construct a multi-dimensional discrimination model for surrounding rock destruction, combines time evolution data to dynamically capture the critical state, and uses a nonlinear dynamic algorithm to predict the development trend of destruction. Through multi-dimensional data analysis and model iterative optimization, the present invention realizes the accurate simulation and prediction of the surrounding rock destruction process under complex geological conditions, providing important technical support for engineering safety. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] The accompanying drawings, which constitute part of this application, are intended to provide a further understanding of this application. The exemplary embodiments and descriptions of this application are intended to explain this application and do not constitute an improper limitation on this application. In the accompanying drawings:

[0018] Figure 1 is a flow chart of a method according to an embodiment of the present invention;

[0019] Figure 2 is a prediction accuracy curve diagram of an embodiment of the present invention;

[0020] Figure 3 A schematic diagram of an evaluation image according to an embodiment of the present invention;

[0021] Figure 4 This is a schematic diagram of an image showing a development trend of predicted surrounding rock failure according to an embodiment of the present invention;

[0022] Figure 5 Schematic diagram of a system according to an embodiment of the present invention. DETAILED DESCRIPTION

[0023] It should be noted that, in the absence of conflict, the embodiments and features of the embodiments in this application can be combined with each other. The present application will be described in detail below with reference to the accompanying drawings and in combination with the embodiments.

[0024] It should be noted that the steps shown in the flowcharts of the accompanying drawings can be executed in a computer system such as a set of computer-executable instructions, and that, although a logical order is shown in the flowcharts, in some cases, the steps shown or described can be executed in an order different from that shown here.

[0025] Example 1

[0026] like Figure 1As shown, this embodiment provides a method for determining damage to water-bearing surrounding rock in underground engineering, including:

[0027] Step S101: obtaining water pressure variation data of water-bearing surrounding rocks, and calculating the variation trend of the surrounding rock stress field in combination with a preset stress distribution model.

[0028] Obtain water pressure variation data for the water-bearing surrounding rock and extract the water pressure variation and variation coefficient. Based on this variation and variation coefficient, combined with a pre-set stress distribution model, extract the stress modulus and distribution state. Calculate the changing trend of the surrounding rock stress field using pre-established calculation methods.

[0029] Specifically, obtaining data on water pressure changes in the water-bearing surrounding rock is a crucial foundation for project safety assessment. For example, at a hydropower station, a network of pressure sensors embedded in the surrounding rock provides real-time monitoring of water pressure. These sensors are positioned at varying depths and locations, forming a three-dimensional monitoring network. Monitoring data show that when the reservoir water level drops from normal storage to dead water, the water pressure in the shallow surrounding rock decreases from 0.2 MPa to 0.1 MPa, while in deeper areas, it decreases from 0.5 MPa to 0.3 MPa. Water pressure changes can cause stress redistribution in the surrounding rock. Combined with a pre-defined stress distribution model, the stress change pattern can be calculated. In this hydropower station case, an elastic-plastic mechanics model was employed, accounting for the heterogeneity and anisotropy of the surrounding rock mass. When monitoring water pressure changes, the stress distribution exhibits distinct nonlinear characteristics, manifesting as stress release in shallow areas and stress concentration in deeper areas.

[0030] An iterative method is used to calculate the trend of the surrounding rock stress field. When the stress in a certain area of the surrounding rock exceeds a preset critical value, the model parameters need to be adjusted. For example, when the stress value at a certain measuring point reaches 70% of the rock mass compressive strength, a warning value is triggered, and the stress distribution model parameters need to be re-optimized, including key parameters such as the rock mass deformation modulus and Poisson's ratio. The updated stress distribution model more accurately reflects the actual state of the surrounding rock. In actual projects, long-term monitoring has shown that the error between the predicted values and the measured values after the adjusted model is within 5%, significantly improving prediction accuracy. This dynamic update mechanism ensures the adaptability and accuracy of the model. Data on the dynamic changes in the surrounding rock stress field are obtained through time series analysis. In this project, six months of continuous monitoring recorded the evolution of the stress field caused by water level changes. The data showed that when the water level drops rapidly, the surrounding rock stress field exhibits a significant hysteresis effect, and the stress adjustment takes approximately three to five days to reach a new equilibrium state. This dynamic monitoring and analysis method can effectively predict the stability of the surrounding rock. By establishing a correlation between water pressure changes and stress field responses, abnormal changes can be detected promptly. For example, if the stress change rate in a certain area exceeds 0.05 MPa per day, it can be identified as an abnormal state, and protective measures such as reinforcement can be taken promptly. This prediction and protection mechanism provides a reliable guarantee for the safe operation of the project.

[0031] Step S102 : Based on the changing trend of the stress field, a crack expansion model is used to simulate the expansion process and direction of microcracks in the surrounding rock.

[0032] Obtain the stress field distribution of the surrounding rock mass, extract the stress values and their changing trends. Use a crack model, combined with the stress value changing trends, to calculate the direction of microcrack expansion. Through simulation, analyze the expansion process of microcracks in the surrounding rock mass.

[0033] Specifically, the stress field distribution of a surrounding rock mass typically manifests as a spatial distribution pattern of normal and shear stresses, with the changing trend of stress values reflecting the dynamic characteristics of the surrounding rock's stress state. For example, in the surrounding rock of a coal mine tunnel, when the vertical stress in the roof reaches 10 MPa, a typical stress concentration zone forms within the surrounding rock mass. This stress concentration often becomes the primary trigger for the initiation of microcracks. The development of a fracture model requires consideration of the surrounding rock's lithologic characteristics and initial fracture distribution. For example, sandstone surrounding rock has a uniaxial compressive strength of approximately 50 MPa. When the stress exceeds a critical value, microcracks begin to propagate. Observations have shown that when vertical stress dominates, microcracks tend to propagate along the direction of maximum principal stress, forming groups of nearly parallel fractures. Simulating and analyzing the microcrack propagation process requires incorporating the physical and mechanical parameters of the surrounding rock. For example, shale surrounding rock has an elastic modulus of approximately 20 GPa and a Poisson's ratio of 0.3. When microcracks propagate, unstable propagation begins when the stress intensity factor at the fracture tip reaches the critical value (square root of 1.5 MPa-m). Adjustment of fracture model parameters primarily considers the anisotropic characteristics of the rock mass. In actual mine applications, when microcrack propagation directions deviate by more than 15 degrees from predicted values, the bedding orientation and joint development of the rock mass need to be reassessed, and the strength parameters in the fracture propagation criterion adjusted accordingly. The dynamic evolution of microcrack propagation can be understood through time-series analysis. For example, within 24 hours after excavation, stress redistribution in the surrounding rock caused rapid microcrack propagation, with crack length growth rates reaching 0.5 mm per hour. Later, the growth rate stabilized, falling to less than 0.1 mm per hour. Key changes in the microcrack propagation process include three stages: crack initiation, accelerated propagation, and stabilization. During the initiation stage, microcracks less than 1 mm in length develop within the surrounding rock. This then enters a phase of accelerated propagation, rapidly extending to the centimeter scale, before ultimately stabilizing after stress redistribution. These changes directly impact the overall stability of the surrounding rock. When multiple fractures connect, a continuous fracture surface may form, leading to rock instability. The deformation characteristics of the surrounding rock mass are closely related to the expansion of microcracks. Displacement monitoring reveals that when the microcrack expansion rate reaches a critical value, the surrounding rock deformation rate increases significantly, with deformation reaching millimeter levels. This change often indicates that the surrounding rock is about to enter a nonlinear deformation stage. Therefore, by analyzing the microcrack expansion pattern, the stability of the surrounding rock can be predicted, providing a basis for support design.

[0034] Step S103 : Based on the simulation results of the crack propagation process and in combination with the physical property parameters of the surrounding rock, the nonlinear characteristics of the crack propagation are determined.

[0035] Through time series analysis, dynamic data on crack expansion is generated and key change points are extracted. Combined with the physical properties of the surrounding rock, the nonlinear characteristic values of crack expansion are analyzed. Whether the characteristic value of crack expansion reaches a preset threshold is determined to determine the stability of the surrounding rock mass.

[0036] Specifically, the stress distribution data for the surrounding rock mass often needs to consider factors such as excavation disturbance and geostress conditions. For example, in a deep-shaft mining project, strain sensors were deployed to measure the principal stress distribution in the surrounding rock mass. The principal stress values gradually decayed from the shaft wall inward, indicating a gradual reduction in the impact of excavation disturbance. The stress values exhibited an exponential decay trend, with the maximum principal stress at the shaft wall reaching more than three times the initial geostress. The selection of a crack propagation model requires a comprehensive consideration of rock mass properties and geological conditions. For a tunnel project, a prediction model for surrounding rock crack propagation was developed based on elastic-plastic mechanics. This model correlates the crack propagation direction with the direction of maximum principal stress and incorporates rock mass strength parameters as control factors. Validation with field data showed that the predicted crack propagation direction agreed within 15% of field observations. In terms of time series analysis, fiber optic sensing was used to monitor the development of surrounding rock cracks in the underground powerhouse of a hydropower station. Analysis of six consecutive months of monitoring data revealed a key change point in the third month after excavation, with a significant increase in the crack propagation rate. This change is related to the seasonal rainfall period, indicating that groundwater seepage has an important influence on crack expansion. During the parameter set adjustment process, the anisotropic characteristics of the surrounding rock mass need to be considered. Monitoring of the surrounding rock in a mine goaf showed that the deviation between the original crack expansion model prediction direction and the actual observation reached 30 degrees. By introducing the bedding angle parameter and adjusting the elastic modulus ratio, the prediction accuracy of the revised model was improved to 95%, such as Figure 2 As shown. The influence of the physical properties of the surrounding rock on the expansion of cracks shows nonlinear characteristics. Taking a deep tunnel project as an example, the uniaxial compressive strength of the surrounding rock mass is 80 MPa. When the stress exceeds the critical value, the crack expansion rate increases exponentially. Measured data show that when the deformation of the surrounding rock exceeds the critical value of 0.3%, the microcracks begin to expand rapidly. A reasonable evaluation index system needs to be established to judge the stability state. In the surrounding rock monitoring of a subway section tunnel, multiple early warning indicators such as displacement and crack width were set. When the displacement rate exceeds 0.5 mm per day or the crack width exceeds 2 mm, it is judged to enter the early warning state. Combined with the numerical simulation results, the critical state threshold of the surrounding rock mass instability was determined, which provides an important reference basis for engineering construction.

[0037] Step S104: obtaining time series data of water pressure fluctuations and analyzing the influence of water pressure on the permeability characteristics of surrounding rocks.

[0038] Specifically, collecting time-series data on water pressure fluctuations is fundamental to studying surrounding rock stability. High-precision pressure sensors can be deployed to obtain water pressure values at different depths. The changing trend of water pressure reflects the dynamic changes in the seepage field within the surrounding rock. For example, a rapid increase in water pressure to 8 MPa over a short period of time indicates that the seepage channel may have partially broken through. Permeability models must consider factors such as the porosity and degree of fracture development in the surrounding rock. For example, in an underground project, three groups of fractures developed in the surrounding rock: the primary fracture had a dip of 65 degrees, and the secondary fractures had dips of 40 and 75 degrees, respectively. The model calculated an equivalent permeability coefficient of 0.003 cm / s. Identifying key change points in time series analysis is crucial for predicting surrounding rock stability. For example, monitoring data showed that the water pressure at a certain point increased from 2 MPa to 6 MPa within 15 days, with an exponential increase. This sudden change often indicates a significant change in the surrounding rock permeability. When measured permeability values exceed the preset safety range, model parameters must be adjusted promptly. For example, in a deep tunnel project, the surrounding rock permeability coefficient exceeded the warning value by 30%. By optimizing the fracture connectivity parameters, the model calculated value was made more consistent with the actual situation. The nonlinearity of permeability characteristics is mainly reflected in the coupling effect between the stress field and the seepage field.

[0039] In actual projects, as groundwater pressure increases, cracks in the surrounding rock gradually open, and the permeability coefficient increases nonlinearly. When the water pressure reaches a critical value, the permeability coefficient may jump. The physical property parameters of the surrounding rock include elastic modulus, Poisson's ratio, etc. These parameters are closely related to the permeability characteristics. The elastic modulus of the surrounding rock in a certain mining area is 25 GPa, and the Poisson's ratio is 0.28. When the surrounding rock is subjected to high water pressure, its permeability coefficient increases exponentially with the deformation of the surrounding rock. Stability assessment requires comprehensive consideration of multiple characteristic values. For example, when the permeability coefficient of the surrounding rock of a certain project exceeds 0.005 cm per second and the deformation reaches the warning value, it indicates that the surrounding rock has entered an unstable state and reinforcement measures are required. The results of the permeability characteristic analysis can provide an important basis for engineering design and construction, ensuring the long-term stability of underground projects.

[0040] Step S105: According to the change of the permeability characteristics, the parameters of the crack expansion model are modified, and the dynamic process of the crack expansion is recalculated.

[0041] Obtain time series data on permeability characteristics and extract trends. Based on these trends, use a pre-defined fracture propagation model and modify its parameters. Using these modified parameters, recalculate the dynamics of fracture propagation and obtain the evolution of fracture propagation.

[0042] Specifically, time series data on permeability characteristics typically reflect the fluid flow state within the surrounding rock mass, recording changes in the permeability coefficient or permeability at fixed time intervals. For example, in a water conservancy project, the permeability coefficient monitored gradually increased from an initial value of 0.05 cm / s to 0.15 cm / s. This trend suggests a possible change in the surrounding rock structure. Fracture propagation models are theoretical tools for describing the development of cracks within a rock mass, often based on elastoplastic or damage models. For example, during the construction of an underground cavern, after an initial crack appeared in the surrounding rock, displacement monitoring data revealed that the crack width expanded from 0.2 mm to 1.5 mm. This necessitated adjustments to the model parameters, including adjustments to key parameters such as the rock mass elastic modulus and Poisson's ratio. The dynamic process of crack propagation involves stages such as crack initiation, expansion, and breakthrough. For example, in a tunnel project, cracks in the surrounding rock gradually evolved from the micron level to the millimeter level, ultimately forming a connected fracture network. This process requires real-time tracking of changes in characteristic parameters such as crack length, aperture, and direction. Adjustment of model parameters is an iterative process based on measured data and theoretical calculations. For example, in a mining project, it was discovered that the initially set rock mass strength parameters deviated from actual monitoring results. Model parameters needed to be optimized repeatedly based on field stress and strain data until the calculated results and measured data were within acceptable error limits. The nonlinear characteristics of crack expansion are reflected in the unevenness of its development rate and direction. For example, during monitoring of the surrounding rock at a hydropower station, it was observed that cracks expanded significantly faster in stress-concentrated areas than in other areas, and the direction of expansion exhibited a complex nonlinear relationship with the principal stress direction. This nonlinear characteristic can be quantified using metrics such as crack density and connectivity. Determining the final state requires comprehensive consideration of multiple factors. In a deep mining project, multiple thresholds for determining stability were established by monitoring parameters such as crack expansion rate and seepage pressure changes. When the crack expansion rate is less than 0.01 mm per day and the seepage pressure fluctuation is less than 5%, the surrounding rock is considered relatively stable. This multi-parameter threshold determination method enables a more accurate assessment of the overall stability of the surrounding rock.

[0043] Step S106 : Based on the corrected dynamic process of crack expansion, a multi-field coupling analysis method is used to construct a multi-dimensional discrimination model of surrounding rock failure.

[0044] The corrected dynamic process data of crack expansion is obtained, and the multi-field coupling analysis method is used to extract the characteristic indicators of surrounding rock failure. Based on the characteristic indicators of surrounding rock failure, a multidimensional discriminant model is constructed to determine the key characteristics of surrounding rock failure. According to the key characteristics of surrounding rock failure, the support vector machine algorithm is used to train the multidimensional discriminant model. The critical state of surrounding rock failure is judged by the trained multidimensional discriminant model. If the critical state of surrounding rock failure reaches the preset threshold, the random forest algorithm is used to optimize the parameters of the multidimensional discriminant model. Based on the optimized model parameters, the multidimensional discriminant model is updated to obtain the discrimination results of surrounding rock failure. Combined with the discrimination results of surrounding rock failure, the gradient boosting decision tree algorithm is used to analyze the evolution trend of surrounding rock failure.

[0045] Specifically, the multi-field coupling analysis method couples the mechanical, seepage, and temperature fields during the dynamic process of crack expansion. For example, in underground projects, excavation causes a redistribution of the surrounding rock stress field, which in turn changes the groundwater seepage field, thereby affecting the surrounding rock temperature distribution. Through the interaction of physical quantities such as stress and strain, seepage pressure, and temperature gradient, characteristic indicators of surrounding rock failure are extracted, such as the extent of the failure zone, the rate of change of the maximum principal stress, and the trend of permeability. When constructing a multidimensional discriminant model based on these characteristic indicators of surrounding rock failure, the interplay of multiple physical quantities must be considered. For example, during the excavation of a hydropower station's underground powerhouse, key physical quantities included surrounding rock deformation, seepage pressure, and crack aperture. By establishing a multidimensional discriminant model, key characteristic values were identified as surrounding rock deformation reaching 10 cm, seepage pressure exceeding 0.5 MPa, and crack aperture greater than 1 mm. The support vector machine algorithm uses historical engineering case data as training samples when training the multidimensional discriminant model. For example, in a tunnel project, monitoring data on surrounding rock deformation, stress, and seepage were collected to create a training dataset. Through kernel function mapping, low-dimensional features are transformed into a high-dimensional feature space, and the optimal classification hyperplane is found to accurately identify the surrounding rock failure state. The critical state of surrounding rock failure is determined based on the output of a multidimensional discriminant model. When the model output approaches a preset threshold, the surrounding rock is considered critical. For example, in a mining project, a critical state is determined when the support pressure exceeds 20 MPa and the surrounding rock displacement rate exceeds 0.5 mm per day. The random forest algorithm optimizes the parameters of the multidimensional discriminant model by constructing an ensemble learning model of multiple decision trees. For example, in a subway project, a decision tree was constructed using different feature combinations, and a voting mechanism was used to determine the optimal parameter combination. The optimized model more accurately reflects the evolution of surrounding rock failure. The gradient boosting decision tree algorithm analyzes the evolution of surrounding rock failure through iterative training to improve model prediction accuracy. For example, in a goaf project, a model was trained using time-series data of surrounding rock failure to predict the future deformation trend of surrounding rock. When the predicted results indicate a continued increase in the surrounding rock deformation rate and the displacement exceeds the warning value, preventive measures such as reinforcement and support are promptly implemented. By extracting characteristic indicators through multi-field coupling analysis and combining them with machine learning algorithms to construct a discriminant model, we can effectively predict the evolution of surrounding rock failure, providing an important basis for engineering safety early warning and risk prevention. Furthermore, the parameter optimization and continuous updating of the multi-dimensional discriminant model ensure the accuracy and reliability of the model's prediction results.

[0046] Step S107: acquiring time evolution data and combining it with a multi-dimensional discriminant model to dynamically capture the critical state of surrounding rock failure.

[0047] Time series data on the surrounding rock evolution process is acquired and evolution features are extracted. A multidimensional discriminant model is used to dynamically capture the extracted evolution features. If the dynamically captured features meet the preset critical state conditions, the critical state of surrounding rock failure is determined.

[0048] Specifically, during the evolution of the surrounding rock mass, monitoring equipment such as displacement sensors and stress sensors are deployed to acquire time series data such as surrounding rock deformation and stress values. For example, in monitoring the surrounding rock mass of a deep tunnel, data was collected hourly for three consecutive months, capturing the evolution of the surrounding rock roof subsidence from an initial 0 mm to 120 mm. The extracted evolutionary features include dynamic parameters such as deformation rate and acceleration, as well as geometric features such as crack extension length and aperture. A multidimensional discriminant model dynamically captures the surrounding rock mass evolutionary characteristics by constructing a feature space. For example, in a mine tunnel, a feature space encompassing stress, displacement, and crack fields was established. When monitored surrounding rock stress exceeds a critical strength, the deformation rate increases suddenly, and crack extension accelerates, the surrounding rock mass is determined to have entered a critical failure state. Specifically, this is manifested by a sudden increase in the roof subsidence rate from 0.5 mm per day to 3 mm per day, and the surrounding rock stress exceeds 80% of the rock mass's compressive strength. A support vector machine algorithm optimizes the discriminant model parameters by finding the optimal classification hyperplane. For example, in monitoring the surrounding rock of a mine tunnel, characteristic parameters such as stress, deformation, and cracks were selected to construct training samples. Kernel function mapping was used to transform the nonlinear problem into a linearly separable one, enabling accurate identification of the surrounding rock failure state. The optimized model improved classification accuracy from 85% to 92%. The gradient boosting decision tree algorithm analyzes the evolutionary trends of surrounding rock failure by constructing a sequence of weak learners. Using a mine tunnel as an example, surrounding rock characteristics such as deformation, stress, and cracks were used as input variables to predict the development trend of surrounding rock failure. The analysis results show that when the roof subsidence rate exceeds 3 mm per day and the surrounding rock stress continues to increase and exceeds the critical strength, the surrounding rock will experience large-scale failure within seven to ten days. Determining the final state through evolutionary trend analysis requires comprehensive consideration of multiple factors. For example, after the surrounding rock of a mine tunnel undergoes stages of initial damage, crack expansion, and critical instability, it may eventually experience large-scale roof collapse and severe deformation of the two sides. Specifically, the roof has sunk by more than 200 mm, the two sides have deformed by more than 300 mm, and through-going cracks are visible on the surface of the surrounding rock. These characteristics collectively indicate that the surrounding rock has entered a state of irreversible damage.

[0049] Step S108: predicting the development trend of surrounding rock failure using a nonlinear dynamics algorithm based on the changing trend of the critical state.

[0050] Obtain the time series data of the surrounding rock mass and extract the characteristic values during the evolution of the surrounding rock. According to the changing trend of the characteristic values, a prediction model is established using a nonlinear dynamic algorithm. The prediction model is used to analyze the development trend of surrounding rock failure, such as Figure 4 shown.

[0051] Specifically, time series data of the surrounding rock mass primarily consists of displacement changes, stress changes, and acoustic emission data, collected in real time via sensors. Taking tunnel construction as an example, displacement data of the surrounding rock mass can be continuously monitored using multi-point displacement meters, with data recorded every hour to form a time series dataset. During feature value extraction, characteristics such as displacement increment, displacement rate, and acceleration can be analyzed, while also incorporating the changing patterns of stress data. In nonlinear dynamic analysis, phase space reconstruction techniques can be used to establish predictive models. For example, in a tunnel project, by embedding dimensionality and time delay parameters, a one-dimensional time series is reconstructed into a multidimensional phase space, revealing the dynamic characteristics of the surrounding rock evolution process. When the surrounding rock mass enters a critical state, the displacement curve exhibits distinct nonlinear characteristics, such as a sudden change point or an accelerated development phase. Multiple indicators must be considered comprehensively when determining the critical state of surrounding rock failure. For example, in a mine tunnel, a critical state can be determined when the surrounding rock displacement rate exceeds 0.5 mm per hour for more than four hours, accompanied by a large number of microseismic signals. During the optimization process of the support vector machine algorithm, monitoring data such as displacement and stress are used as training samples, mapped to high-dimensional feature space through kernel functions, and a classification hyperplane is established to achieve accurate identification of critical states. The update of the prediction model is mainly based on the newly added monitoring data. Taking an underground powerhouse as an example, when the new data shows that the deformation rate of the surrounding rock has increased significantly, the prediction model needs to adjust the weights of relevant parameters in a timely manner to improve the prediction accuracy. Trend analysis can predict the development of deformation in a certain period in the future and provide a decision-making basis for engineering disposal. When the gradient boosting decision tree algorithm is used to verify the model, the historical data can be divided into a training set and a test set. Taking a deep well mining site as an example, 70% of the data is selected as the training set, and the remaining data is used for testing. The prediction ability of the model is evaluated by comparing the error between the predicted value and the measured value. When the prediction accuracy reaches more than 85% and the discreteness of the prediction results is small, it means that the model has good stability. Figure 3 The figure shows a comparison of the accuracy, precision, and recall of the support vector machine algorithm, random forest algorithm, and gradient boosting decision tree algorithm. The gradient boosting decision tree algorithm performed best in all three metrics, achieving an accuracy of 95%. This model can be used to guide engineering practice, promptly identify potential surrounding rock instability risks, and provide a scientific basis for proactive prevention and control.

[0052] Example 2

[0053] like Figure 5 As shown, based on the same inventive concept, this embodiment also provides a system for determining damage to water-bearing surrounding rock in underground engineering, the system comprising:

[0054] The water pressure change data acquisition module is used to obtain the water pressure change data of the water-bearing surrounding rock and calculate the change trend of the surrounding rock stress field in combination with the preset stress distribution model;

[0055] The stress field calculation module is used to simulate the expansion process and direction of microcracks in the surrounding rock using a crack expansion model based on the changing trend of the stress field;

[0056] The crack propagation simulation module is used to determine the nonlinear characteristics of crack propagation based on the simulation results of the crack propagation process and the physical properties of the surrounding rock;

[0057] Nonlinear characteristic judgment module, used to obtain time series data of water pressure fluctuations and analyze the influence of water pressure on the permeability characteristics of surrounding rocks;

[0058] The water pressure fluctuation analysis module is used to modify the parameters of the fracture propagation model according to the changes in permeability characteristics and recalculate the dynamic process of fracture propagation;

[0059] The crack extension model correction module is used to construct a multi-dimensional discrimination model of surrounding rock failure based on the corrected crack extension dynamic process and the multi-field coupling analysis method;

[0060] Multidimensional discriminant model building module, used to obtain time evolution data and dynamically capture the critical state of surrounding rock failure in combination with multidimensional discriminant models;

[0061] The critical state capture module is used to predict the development trend of surrounding rock failure based on the changing trend of the critical state using nonlinear dynamics algorithm.

[0062] The system for distinguishing the damage of water-bearing surrounding rock in underground engineering provided by this embodiment has all the advantages of the method for distinguishing the damage of water-bearing surrounding rock in underground engineering provided by the first embodiment.

[0063] Example 3

[0064] This embodiment further discloses a computer program product, including a computer program, which implements the steps of the method described in the first embodiment when executed by a processor.

[0065] The above are merely preferred embodiments of the present application, but the scope of protection of the present application is not limited thereto. Any changes or substitutions that can be easily conceived by a person skilled in the art within the technical scope disclosed in this application should be included in the scope of protection of the present application. Therefore, the scope of protection of the present application should be based on the scope of protection of the claims.

Claims

1. A method for determining the damage of water-bearing surrounding rock in underground engineering, characterized in that: The following steps are involved: Obtain water pressure change data of water-bearing surrounding rocks and calculate the change trend of surrounding rock stress field by combining with the preset stress distribution model; According to the change trend of the surrounding rock stress field, a crack expansion model is used to simulate the expansion process and direction of microcracks in the surrounding rock; Based on the simulation results of the microcrack propagation process and combined with the physical properties of the surrounding rock, the nonlinear characteristics of crack propagation are determined; Obtain time series data of water pressure fluctuations and analyze the influence of water pressure on the permeability characteristics of surrounding rocks; According to the change of the permeability characteristics, the parameters of the crack expansion model are modified to recalculate the dynamic process of the crack expansion; Based on the modified dynamic process of crack expansion, a multi-dimensional discrimination model of surrounding rock failure is constructed using multi-field coupling analysis method. Acquire time evolution data and combine it with the multidimensional discriminant model to dynamically capture the critical state of surrounding rock failure; According to the changing trend of the critical state, a nonlinear dynamics algorithm is used to predict the development trend of surrounding rock failure.

2. The method according to claim 1, characterized in that The steps for calculating the changing trend of the surrounding rock stress field include: Obtain water pressure change data of water-bearing surrounding rocks and extract water pressure change values; Extracting the stress modulus and distribution state based on the water pressure change value and a preset stress distribution model; The changing trend of the surrounding rock stress field is calculated using a pre-established calculation method.

3. The method according to claim 1, characterized in that The steps to simulate the propagation process and direction of microcracks in the surrounding rock include: Obtain the stress field distribution state of the surrounding rock mass, extract the stress value and its change trend; Using a crack model and combining the stress value change trend, the expansion direction of the microcracks is calculated; Through the simulation process, the expansion process of microcracks in the surrounding rock mass is analyzed.

4. The method according to claim 1, wherein The steps for determining the nonlinear characteristics of crack expansion include: Through time series analysis, dynamic data of crack expansion is generated and key change points are extracted; Combined with the physical properties of the surrounding rock, the nonlinear characteristic values of crack expansion are analyzed; It is judged whether the crack expansion characteristic value reaches a preset threshold value to determine the stability state of the surrounding rock mass.

5. The method according to claim 1, wherein The steps to recalculate the dynamics of crack growth include: Obtain time series data of permeability characteristics and extract the changing trend of permeability characteristics; According to the changing trend of permeability characteristics, the preset crack expansion model is used to modify the model parameter values; According to the revised parameter values, the dynamic process of crack expansion is recalculated to obtain the evolution results of crack expansion.

6. The method according to claim 1, characterized in that The steps of constructing a multi-dimensional discrimination model of surrounding rock failure using the multi-field coupling analysis method include: Obtain the corrected dynamic process data of crack expansion and use multi-field coupling analysis method to extract characteristic indicators of surrounding rock failure; Based on the characteristic indicators of surrounding rock failure, a multi-dimensional discrimination model is constructed to determine the key characteristics of surrounding rock failure; According to the key characteristics of surrounding rock failure, the support vector machine algorithm is used to train the multidimensional discriminant model to obtain the trained multidimensional discriminant model.

7. The method according to claim 1, characterized in that The steps to dynamically capture the critical state of surrounding rock failure include: Obtain time series data of surrounding rock evolution and extract evolution characteristics; A multi-dimensional discriminant model is used to dynamically capture the extracted evolution features; If the dynamically captured features meet the preset critical state conditions, the critical state of surrounding rock failure is determined.

8. The method according to claim 1, characterized in that The steps to predict the development trend of surrounding rock failure include: Obtain time series data of surrounding rock mass and extract characteristic values during the evolution of surrounding rock mass; According to the changing trend of the characteristic values, a nonlinear dynamics algorithm is used to establish a prediction model; The development trend of surrounding rock failure is analyzed through prediction models.

9. A system for determining damage to surrounding rock in underground engineering, characterized by: The system comprises: The water pressure change data acquisition module is used to obtain the water pressure change data of the water-bearing surrounding rock and calculate the change trend of the surrounding rock stress field in combination with the preset stress distribution model; The stress field calculation module is used to simulate the expansion process and direction of microcracks in the surrounding rock using a crack expansion model based on the changing trend of the stress field; The crack propagation simulation module is used to determine the nonlinear characteristics of crack propagation based on the simulation results of the crack propagation process and the physical properties of the surrounding rock; Nonlinear characteristic judgment module, used to obtain time series data of water pressure fluctuations and analyze the influence of water pressure on the permeability characteristics of surrounding rocks; The water pressure fluctuation analysis module is used to modify the parameters of the fracture propagation model according to the changes in permeability characteristics and recalculate the dynamic process of fracture propagation; The crack extension model correction module is used to construct a multi-dimensional discrimination model of surrounding rock failure based on the corrected crack extension dynamic process and the multi-field coupling analysis method; Multidimensional discriminant model building module, used to obtain time evolution data and dynamically capture the critical state of surrounding rock failure in combination with multidimensional discriminant models; The critical state capture module is used to predict the development trend of surrounding rock failure based on the changing trend of the critical state using nonlinear dynamics algorithm.

10. A computer program product comprising a computer program, characterized in that When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 8 are implemented.

Images