Optical fiber sensing-based strip mine monitoring method, device, equipment and medium

Abstract

The invention discloses an open pit mine monitoring method, device and equipment based on optical fiber sensing and a medium, and belongs to the technical field of open pit coal mine slope monitoring, and the method comprises the following steps: obtaining geological parameters and environmental parameters of a to-be-monitored area; performing multi-physical field simulation according to the geological parameters to obtain simulation sensing data; carrying out random capacity expansion on the environment parameters to obtain capacity expansion data; according to the capacity expansion data, a limit equilibrium algorithm is adopted to calculate a safety coefficient; inputting the simulated sensing data and the safety coefficient as training samples into a neural network model for model training to obtain a monitoring model; and acquiring real-time sensing data, and inputting the real-time sensing data into the monitoring model to obtain a monitoring result. According to the method, multi-physics field simulation is carried out by obtaining geological parameters to obtain simulation sensing data for model training, then monitoring is carried out through the model, the slope stability is increased from a traditional single physics field to multi-physics field coupling, and the prediction accuracy is larger than or equal to 95%.

Description

Technical Field

[0001] This invention belongs to the field of mine monitoring technology, and particularly relates to a method, device, equipment and medium for monitoring open-pit mines based on fiber optic sensing. Background Technology

[0002] Traditional open-pit mine monitoring often relies on single physical field data (such as monitoring only displacement or stress), ignoring the coupling effect of multiple physical fields such as stress-displacement-seepage, resulting in one-sided monitoring results that are difficult to reflect the true geological disaster risks.

[0003] The Limit Equilibrium (LEM) method is commonly used for slope stability analysis. However, this method is cumbersome to calculate, relies on artificially assumed slip surfaces, has large subjective errors, and does not fully consider the interaction between the support structure and the geological body, making it unsuitable for complex open-pit mining conditions.

[0004] Although fiber optic sensing technology has been applied to mine monitoring, data processing mostly uses traditional statistical methods and lacks intelligent analysis models, resulting in low processing efficiency and delayed early warning. At the same time, existing software simulations mostly focus on a single physical field and have not achieved linkage analysis with sensor data.

[0005] The application of intelligent algorithms in slope monitoring is limited to small sample training. Insufficient sample diversity leads to poor model generalization ability and makes it difficult to adapt to the dynamic changes of complex geological conditions in open-pit mines. Summary of the Invention

[0006] In view of the technical problems existing in the prior art, the present invention provides a method, device, equipment and medium for monitoring open-pit mines based on optical fiber sensing.

[0007] According to a first aspect of the technical solution of the present invention, a monitoring method for open-pit mines based on optical fiber sensing is provided, which includes the following steps: S1: Obtain geological and environmental parameters of the area to be monitored; S2: Multiphysics simulation is performed based on the geological parameters obtained in S1 to obtain simulated sensor data; S3: Randomly expand the environmental parameters obtained in S1 to obtain expanded data; S4: The safety factor is calculated using the limit balance algorithm based on the expansion data obtained in S3; S5: Use the simulated sensing data obtained in S2 and the safety coefficient obtained in S4 as training samples to input into the neural network model for model training, and obtain the monitoring model; S6: Acquire real-time sensing data and input the real-time sensing data into the monitoring model acquired in S5 to obtain the monitoring results.

[0008] A further improvement of the present invention is that step S2 includes the following steps: S21: Establish a multiphysics 2D model based on the geological parameters obtained in S1; S22: The first potential slip surface is obtained by using the strength reduction method based on the multiphysics 2D model obtained in S21; S23: Read the simulated sensing data of the first potential slip surface obtained in S22 based on the multiphysics 2D model obtained in S21.

[0009] A further improvement of the present invention is that the geological parameters include rock mass density, elastic modulus, internal friction angle and cohesion, and the environmental parameters include physical parameters of the soil and rock mass, geometric parameters of the slope, seepage environment parameters and support structure parameters.

[0010] A further improvement of the present invention is that, in S3, the expanded data is obtained by randomly adjusting the environmental parameters according to a preset range.

[0011] A further improvement of the present invention is that step S4 includes the following steps: S41: The second potential slip surface is obtained by using the strength reduction method based on the expansion data obtained in S3; S42: Divide the second potential slip surface obtained in S41 into several soil strips along the vertical direction; S43: Calculate the bottom slip surface length and weight of each soil strip obtained in S42 based on the expansion data obtained in S3; S44: The anti-sliding force and tangential force of each soil strip are calculated based on the force balance analysis, the bottom slip surface length of each soil strip obtained in S43, and the weight of each soil strip obtained in S43. S45: The safety factor is calculated based on the anti-sliding force and tangential force of each soil strip obtained in S44.

[0012] A further improvement of the present invention is that the monitoring model includes: The input layer is used to receive the analog sensing data obtained in S2 and the safety factor obtained in S4; Convolutional layers are used to obtain the local spatial correlation between the simulated sensing data and the safety factor; The time series analysis layer is used to predict the slope stability change trend based on the simulated sensor data and the safety factor, and obtain the stability change trend. The fully connected layer includes two hidden layers, each containing an activation function to dynamically adjust the weights of the simulated sensing data and the safety coefficient, and to fuse the trend vector T. The output layer is used to output monitoring results, which include the current security factor and the security level corresponding to the current security factor.

[0013] A further improvement of the present invention is that step S6 includes the following steps: S61: Acquire real-time sensing data and input the real-time sensing data into the monitoring model acquired in S5 to obtain the prediction result; S62: Obtain stress sensor data and ultimate bending moment of the mine support structure; S63: Calculate the maximum bending moment of the mine support structure based on the stress sensor data obtained in S62; S64: The monitoring results are calculated based on the ultimate bending moment obtained in S62, the maximum bending moment obtained in S63, and the prediction results obtained in S61.

[0014] According to a second aspect of the technical solution of the present invention, an open-pit mine monitoring device based on optical fiber sensing is provided, which employs the above-mentioned open-pit mine monitoring method based on optical fiber sensing, and includes: The data acquisition module is used to acquire geological and environmental parameters of the area to be monitored. The simulation module is used to perform multiphysics field simulations based on the geological parameters to obtain simulated sensor data; The expansion module is used to randomly expand the environmental parameters to obtain expanded data. The safety factor calculation module is used to calculate the safety factor based on the expansion data using a limit balance algorithm. The training module is used to input the simulated sensing data and the safety coefficient as training samples into the neural network model to train the model and obtain the monitoring model. The prediction module is used to acquire real-time sensor data and input the real-time sensor data into the monitoring model to obtain monitoring results.

[0015] According to a third aspect of the technical solution of the present invention, a computer device is provided, including a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that the processor executes the computer program to implement the above-mentioned open-pit mine monitoring method based on fiber optic sensing.

[0016] According to a fourth aspect of the present invention, a computer-readable storage medium is provided, on which a computer program is stored, wherein the computer program, when executed by a processor, implements the above-described open-pit mine monitoring method based on fiber optic sensing.

[0017] This invention obtains simulated sensor data by acquiring geological parameters and performing multi-physics field simulation. Based on the simulated sensor data, the model is trained and then monitored through the model. This achieves a leap from traditional single-physics field, small-sample, qualitative judgment to multi-physics field coupling, large-sample, quantitative intelligent prediction of slope stability, making the prediction accuracy ≥95% and meeting the 24-hour early warning requirement of open-pit mines.

[0018] This invention achieves precise binding between the location of the slip surface and the strain-temperature-displacement response of the optical fiber by identifying the slip surface and then obtaining simulated sensing data. This provides high-fidelity training samples with "damage feature labels" for the subsequent neural network, controls the slip surface identification error to ≤5%, and significantly improves the model's ability to reproduce real damage modes.

[0019] This invention improves upon the traditional extreme balance method by 80% through a CNN architecture and a neural network structure without pooling layers, with a single data processing time of ≤0.5 seconds, meeting the requirements for real-time monitoring.

[0020] The method of this invention effectively reduces the cost of manual exploration and data analysis by more than 60%, reduces economic losses caused by open-pit mine landslides, and enhances the mine's safety production guarantee capabilities. Attached Figure Description

[0021] The accompanying drawings are provided to better understand the invention and are not intended to unduly limit the scope of the invention. Wherein: Figure 1 This is a flowchart of an open-pit mine monitoring method based on fiber optic sensing according to the present invention; Figure 2 This is a structural block diagram of an open-pit mine monitoring device based on fiber optic sensing according to the present invention.

[0022] In the diagram: 800, Computer System; 801, Central Processing Unit; 802, Read-Only Memory; 803, Random Access Memory; 804, Bus; 805, I / O Interface; 806, Input Section; 807, Output Section; 808, Storage Section; 809, Communication Section; 810, Driver; 811, Removable Media. Detailed Implementation

[0023] The following description, in conjunction with the accompanying drawings, illustrates exemplary embodiments of the present invention, including various details to aid understanding. These details should be considered merely exemplary. Therefore, those skilled in the art will recognize that various changes and modifications can be made to the embodiments described herein without departing from the scope and spirit of the invention. Similarly, for clarity and brevity, descriptions of well-known functions and structures are omitted in the following description.

[0024] Example 1 like Figure 1 As shown, this invention provides a monitoring method for open-pit mines based on fiber optic sensing, which includes the following steps: S1: Obtain geological and environmental parameters of the area to be monitored; S2: Multiphysics simulation is performed based on the geological parameters obtained in S1 to obtain simulated sensor data; S3: Randomly expand the environmental parameters obtained in S1 to obtain expanded data; S4: The safety factor is calculated using the limit balance algorithm based on the expansion data obtained in S3; S5: Use the simulated sensing data obtained in S2 and the safety coefficient obtained in S4 as training samples to input into the neural network model for model training, and obtain the monitoring model; S6: Acquire real-time sensing data and input the real-time sensing data into the monitoring model acquired in S5 to obtain the monitoring results.

[0025] Specifically, the geological parameters include parameters such as rock mass density, elastic modulus, internal friction angle, and cohesion, while the environmental parameters include parameters such as physical parameters of the soil and rock mass, geometric parameters of the slope, seepage environment parameters, and support structure parameters.

[0026] Specifically, S2 includes the following steps: S21: Establish a multiphysics 2D model based on the geological parameters obtained in S1; Specifically, the multiphysics 2D model is constructed using finite element method (FEM) software. By inputting the geological parameters, it simulates the coupling process of multiple physics fields such as stress, displacement, and seepage. The multiphysics 2D model is labeled with geological layers (surface landslide body, weak layer, and completely weathered rock layer), fiber optic sensor array (5m spacing, deployed along the slope surface and inside), anti-slide piles (35m in length, 2×2m in cross-section, 4m spacing), and potential slip surfaces (identified by SRM). This figure mainly reflects the distribution cloud map of the multiphysics fields (stress, displacement, and seepage) (with different colors indicating the intensity level), clarifies the positional relationship between the sensors and the geological body and support structure, and shows the model composition and data acquisition source of the multiphysics coupling simulation.

[0027] Specifically, ABAQUS software was used to construct a multiphysics 2D model, setting free boundaries (slope surface / top of slope) and fixed boundaries (bottom) to simulate the coupling process of multiphysics fields such as stress, displacement, and seepage.

[0028] S22: The first potential slip surface is obtained by using the strength reduction method based on the multiphysics 2D model obtained in S21; S23: Read the simulated sensing data of the first potential slip surface obtained in S22 based on the multiphysics 2D model obtained in S21.

[0029] Specifically, the first potential slip surface and the simulated sensor data are obtained using finite element software.

[0030] Specifically, in S3, expanded data is obtained by randomly adjusting the environmental parameters according to a preset range. Expanded data is obtained by randomly adjusting the physical parameters (such as cohesion or friction angle) and geometric parameters (such as slope height or slope angle) within a threshold range of 0-10%. The amount of expanded data is set according to actual needs.

[0031] Specifically, S4 includes the following steps: S41: The second potential slip surface is obtained by using the strength reduction method based on the expansion data obtained in S3; Specifically, the second potential slip surface is obtained through software simulation.

[0032] S42: Divide the second potential slip surface obtained in S41 into several soil strips along the vertical direction; Specifically, the subsequent steps apply to the case where the second potential slip surface is a circular arc slip surface, balancing calculation accuracy and engineering practicality. The core is to derive the ratio of the anti-slip force to the sliding force through the force balance of the soil strips, where the second potential slip surface is a circular arc surface; the sliding body is divided into n soil strips along the vertical direction, with the strip width b preferably being 2m-5m; the tangential force between soil strips is ignored, and only the horizontal force between strips (E) is considered. i ) and vertical inter-strip force (X) i ), and E i =E i+1 (Horizontal force balance); Considering the influence of seepage in open-pit mines, there is pore water pressure u at the bottom of the soil strip. i Effective stress is calculated by subtracting pore water pressure from total stress; physical parameters of each soil layer are taken as simulated / measured values ​​(cohesion c). i internal friction angle φ i Natural bulk density γ i saturated bulk density γ sat,i The parameters of different soil layers are calculated in segments. The slope geometric parameters are as follows (these parameters are for illustrative purposes only and should be obtained according to the actual working conditions): Slope height H = 45m, slope angle θ = 30°, coordinates of the center of the slip surface (x0, y0), radius of the slip surface R; Landslide body (surface layer): c1 = 19kPa, φ1 = 27°, γ1 = 19kN / m 3 γ sat,1 =20kN / m 3 Weak layer (intermediate layer): c2=15.3kPa, φ2=20°, γ2=18kN / m 3 γ sat,2 =19kN / m 3 Completely weathered rock strata (bottom layer): c3=22.5kPa, φ3=37°, γ3=23kN / m 3 Weak layer (intermediate layer): c2=15.3kPa, φ2=20°, γ2=18kN / m3 γ sat,2 =19kN / m 3 Completely weathered rock strata (bottom layer): c3=22.5kPa, φ3=37°, γ3=23kN / m 3 Flow parameters: groundwater level depth h w =10m, pore water pressure coefficient r u =0.3 (Common value for open-pit mine slopes).

[0033] Specifically, in step S42, the length L of the second potential slip surface and the inclination angle α of the slip surface at the bottom of each soil strip are obtained by extracting the coordinates of each point on the arc of the second potential slip surface. i (The angle between the bottom of the i-th soil strip and the horizontal direction).

[0034] S43: Calculate the bottom slip surface length and weight of each soil strip obtained in S42 based on the expansion data obtained in S3; Specifically, in S43, the weight of each soil strip, i.e., the weight W of the i-th soil strip... i The calculation formula is as follows:

[0035]

[0036] In the formula, γ i Let b be the natural unit weight of the soil layer containing the i-th soil strip, b be the width of the soil strip, and h be the weight of the soil strip. w h is the depth of the groundwater level. i Let be the vertical height of the i-th soil strip. γ is the buoyant unit weight of the soil layer containing the i-th soil strip. sat,i γ is the saturated unit weight of the soil layer containing the i-th soil strip. w This is the specific gravity of water.

[0037] The length of the bottom slip surface of each soil strip, i.e., the length l of the bottom slip surface of the i-th soil strip. i The calculation formula is as follows: l i =b / cosα i ; S44: The anti-sliding force and tangential force of each soil strip are calculated based on the force balance analysis, the bottom slip surface length of each soil strip obtained in S43, and the weight of each soil strip obtained in S43. Specifically, the anti-sliding force calculation for each soil strip, i.e., the anti-sliding force T of the i-th soil strip. res，i The calculation formula is as follows:

[0038] In the formula, F SFor the safety factor, c i : Cohesion (kPa) of the soil layer containing the i-th soil strip; : The internal friction angle (°) of the soil layer containing the i-th soil strip; Vertical force balance: W i =N i *cosα i +T i *sinα i Where: N i T represents the effective normal force at the bottom of the soil strip (kN / m). i The actual tangential force at the bottom of the soil strip (kN / m); Relationship between sliding force and tangential force: Actual tangential force T i Determined by the sliding force, the anti-slip force T res,i The safety factor F is provided by the shear strength of the soil. S It is the ratio of total anti-skid force to total tangential force.

[0039] S45: The safety factor is calculated based on the anti-sliding force and tangential force of each soil strip obtained in S44.

[0040] Specifically, in S45, the formulas for vertical equilibrium and anti-slip force are combined, and N is eliminated. i The safety factor is obtained by iterative solution, and the iterative formula is as follows:

[0041] In the formula, W i : Weight of the i-th soil strip (kN / m); α i : The angle (°) between the bottom slip surface of the i-th soil strip and the horizontal direction; l i : Length of the slip surface at the bottom of the i-th soil strip (m).

[0042] Specifically, the iterative solution process is as follows: Initial assumption F S0 =1.0, F S0 Substituting into the right side of the formula, we obtain the first iteration value F. S1 If |F S1 -F S0 If |≤0.01, then F S1 Let F be the final safety factor; otherwise, let F S0 =F S1 Repeat the iteration until convergence; if there are multiple layers of soil and rock, substitute c according to the soil layer where the soil strip is located. i φ i Calculate and ensure parameter matching. Based on the converged F... S Sample labels are assigned according to warning level: Unstable: F S <1.00; Understability: 1.00≤F S<1.05; Basically stable: 1.05≤F S <1.25; Stable: F S ≥1.25.

[0043] Specifically, the structure of the neural network model is the same as that of the monitoring model, with the only difference being the weight coefficients. The monitoring model is obtained by training the neural network model using training data. The monitoring model includes: The input layer is used to receive the analog sensing data obtained in S2 and the safety factor obtained in S4; Convolutional layers are used to obtain the local spatial correlation between the simulated sensing data and the safety factor; The time series analysis layer is used to predict the slope stability change trend based on the simulated sensor data and the safety factor, and obtain the stability change trend. The fully connected layer includes two hidden layers, each containing an activation function to dynamically adjust the weights of the simulated sensing data and the safety coefficient, and to fuse the trend vector T.

[0044] The output layer is used to output monitoring results, which include the current security factor and the security level corresponding to the current security factor.

[0045] Specifically, the neural network model is a four-layer convolutional neural network, employing the Leaky-ReLU activation function without pooling layers to preserve data features. The batch size is set to 30 (which has been verified as the optimal parameter). Simulated sensor data is used as input. The safety factor, the ratio of the anti-sliding force to the sliding force of the open-pit mine slope under current conditions, is a core indicator for quantifying slope stability (unitless), not a model prediction. This safety factor serves as the true label for model training, a quantitative standard for judging whether the slope will become unstable and landslide. The safety factor is related to the physical parameters of the soil and rock mass, the geometric parameters of the slope, the seepage environment parameters, and the parameters of the support structure. The model is trained, and the network weights are optimized using root mean square error (RMSE) and mean relative error (MRE). Through a four-layer CNN architecture, Leaky-ReLU activation, no pooling layers, and a batch size of 30, along with RMSE / MRE-guided weight adjustments, the core capabilities of the CNN network in open-pit mine slope monitoring scenarios are significantly improved, thus meeting monitoring requirements. The neural network model possesses feature extraction capabilities, gradient convergence capabilities, generalization capabilities, and error control capabilities. The feature extraction capability improves recognition accuracy with small sample sizes by preserving detailed features. The gradient convergence capability improves model training efficiency by addressing gradient vanishing. The generalization capability reduces overfitting risk by adapting to diverse working conditions. The error control capability improves quantification accuracy by reducing prediction bias. The output of the actual safety factor is used to solve the quantitative decision-making problem in open-pit mine slope monitoring (replacing qualitative judgment, achieving quantitative stability assessment, adapting to actual engineering needs, and serving as the core triggering condition for early warning logic). The formula for the root mean square error (RMSE) involved is as follows:

[0046] In the formula, N is, F S,pred,i F is the safety coefficient obtained from training the i-th set of samples. S,true,i The safety factor is obtained by using the limit balance algorithm for the i-th group of samples.

[0047] Specifically, the training objective of the neural network model is an RMSE less than or equal to 0.22, meaning that the average deviation between the predicted value and the true value is controlled within 0.22. The formula for calculating the average relative error is as follows:

[0048] Specifically, MRE≤8% means that the average relative deviation between the predicted value and the actual value does not exceed 8%.

[0049] Specifically, step S6 includes the following steps: S61: Acquire real-time sensing data and input the real-time sensing data into the monitoring model acquired in S5 to obtain the prediction result; S62: Obtain stress sensor data and ultimate bending moment of the mine support structure; Specifically, the mine support structure is generally an anti-slide pile. The stress sensor data is obtained by fiber optic grating stress sensors installed at key sections of the anti-slide pile (e.g., near the slip surface, at the top of the pile, at the bottom of the pile, or at 1 / 3 of the pile). The stress sensor data includes the normal stress and shear stress of the pile section, which are used to reflect the magnitude of the stress on the section and determine whether it is close to the ultimate stress. Fiber optic grating strain sensors are deployed along the longitudinal direction of the pile at intervals of 2m-3m, covering the area above and below the slip surface, to measure the strain distribution of the pile.

[0050] S63: Calculate the maximum bending moment of the mine support structure based on the stress sensor data obtained in S62; Specifically, the formula for calculating the bending moment is as follows: M=EIε; In the formula, M: measured bending moment of the anti-slide pile (kN·m), i.e., the bending moment value of the current section; E: elastic modulus of the pile material (kPa or MPa), reflecting the material stiffness; I: moment of inertia of the pile section, which is related to the shape and size of the section; ε: stress sensor data (με).

[0051] Bending moment is calculated using strain gauges; horizontal displacement at the pile top and rotation angle at the pile mid-section (corresponding to the slip surface) are measured using displacement gauges to reflect overall pile displacement and determine if the support effect meets standards; pore water pressure sensors are deployed in the soil around the pile (within a 5m range above and below the slip surface) to measure pore water pressure around the pile and correct for the influence of seepage on pile stress. The maximum bending moment M is calculated. 实测 and ultimate bending moment M 极限 The key ratios are then substituted to obtain the result.

[0052] S64: The monitoring results are calculated based on the ultimate bending moment obtained in S62, the maximum bending moment obtained in S63, and the prediction results obtained in S61.

[0053] Specifically, based on the ultimate bending moment obtained in S62, the maximum bending moment obtained in S63, and the prediction result F obtained in S61... S，预测 The formula for obtaining the monitoring results is as follows: F S,修正 =F S,预测 ×[1-k(M 实测 / M 极限 -0.5)]; In the formula, M 实测 : Maximum bending moment of the anti-slide pile measured by the sensor (kN·m); M 极限: Design ultimate bending moment of anti-slide pile (determined by pile material strength and cross-sectional dimensions); k: Correction coefficient (empirical value 0.3-0.5, 0.4 for open-pit mine / tunnel-landslide systems). F S,修正 The safety factor corresponding to the monitoring results.

[0054] Example 2 like Figure 2 As shown, an open-pit mine monitoring device based on fiber optic sensing is provided, which adopts the open-pit mine monitoring method based on fiber optic sensing in the above embodiments, and includes: The data acquisition module is used to acquire geological and environmental parameters of the area to be monitored. The simulation module is used to perform multiphysics field simulations based on the geological parameters to obtain simulated sensor data; The expansion module is used to randomly expand the environmental parameters to obtain expanded data. The safety factor calculation module is used to calculate the safety factor based on the expansion data using a limit balance algorithm. The training module is used to input the simulated sensing data and the safety coefficient as training samples into the neural network model to train the model and obtain the monitoring model. The prediction module is used to acquire real-time sensor data and input the real-time sensor data into the monitoring model to obtain monitoring results.

[0055] Specifically, the simulation module includes: The model building submodule is used to build a multiphysics 2D model based on the geological parameters. Specifically, the multiphysics 2D model is constructed using finite element software, and the geological parameters are input to simulate the coupling process of multiple physical fields such as stress, displacement, and seepage.

[0056] Specifically, ABAQUS software was used to construct a multiphysics 2D model, setting free boundaries (slope surface / top of slope) and fixed boundaries (bottom) to simulate the coupling process of multiphysics fields such as stress, displacement, and seepage.

[0057] The first identification submodule is used to obtain the first potential slip surface based on the multiphysics 2D model using the strength reduction method. The simulation data acquisition submodule is used to read the simulation sensing data of the first potential slip surface based on the multiphysics 2D model.

[0058] Specifically, the first potential slip surface and the simulated sensor data are obtained using finite element software.

[0059] Specifically, the training module includes: The second identification submodule is used to obtain the second potential slip surface based on the expanded data using the strength reduction method; Specifically, the second potential slip surface is obtained through software simulation.

[0060] A sub-module is used to divide the second potential slip surface into several soil strips along the vertical direction; Specifically, the subsequent steps apply to the case where the second potential slip surface is a circular arc slip surface, balancing calculation accuracy and engineering practicality. The core is to derive the ratio of the anti-slip force to the sliding force through the force balance of the soil strips, where the second potential slip surface is a circular arc surface; the sliding body is divided into n soil strips along the vertical direction, with the strip width b preferably being 2m-5m; the tangential force between soil strips is ignored, and only the horizontal force between strips (E) is considered. i ) and vertical inter-strip force (X) i ), and E i =E i+1 (Horizontal force balance); Considering the influence of seepage in open-pit mines, there is pore water pressure u at the bottom of the soil strip. i Effective stress is calculated by subtracting pore water pressure from total stress; physical parameters of each soil layer are taken as simulated / measured values ​​(cohesion c). i internal friction angle φ i Natural bulk density γ i saturated bulk density γ sat,i The parameters of different soil layers are calculated in segments. The slope geometric parameters are as follows (these parameters are for illustrative purposes only and should be obtained according to the actual working conditions): Slope height H = 45m, slope angle θ = 30°, coordinates of the center of the slip surface (x0, y0), radius of the slip surface R; Landslide body (surface layer): c1 = 19kPa, φ1 = 27°, γ1 = 19kN / m 3 γ sat,1 =20kN / m 3 Weak layer (intermediate layer): c2=15.3kPa, φ2=20°, γ2=18kN / m 3 γ sat,2 =19kN / m 3 Completely weathered rock strata (bottom layer): c3=22.5kPa, φ3=37°, γ3=23kN / m 3 Weak layer (intermediate layer): c2=15.3kPa, φ2=20°, γ2=18kN / m 3 γ sat,2 =19kN / m 3 Completely weathered rock strata (bottom layer): c3=22.5kPa, φ3=37°, γ3=23kN / m 3 Flow parameters: groundwater level depth h w =10m, pore water pressure coefficient r u =0.3 (Common value for open-pit mine slopes).

[0061] Specifically, in the partitioning submodule, the length L of the second potential slip surface and the inclination angle α of the slip surface at the bottom of each soil strip are obtained by extracting the coordinates of each point on the arc of the second potential slip surface. i (The angle between the bottom of the i-th soil strip and the horizontal direction).

[0062] The first soil strip calculation submodule is used to calculate the bottom slip surface length and weight of each soil strip based on the expansion data; Specifically, in the first soil strip calculation submodule, the weight of each soil strip, i.e., the weight W of the i-th soil strip, is... i The calculation formula is as follows:

[0063] The length of the bottom slip surface of each soil strip, i.e., the length l of the bottom slip surface of the i-th soil strip. i The calculation formula is as follows: l i =b / cosα i The second soil strip calculation submodule is used to calculate the anti-sliding force and tangential force of each soil strip based on the force balance analysis, the length of the bottom slip surface of each soil strip, and the weight of each soil strip. Specifically, the anti-sliding force calculation for each soil strip, i.e., the anti-sliding force T of the i-th soil strip. res，i The calculation formula is as follows:

[0064] In the formula, F S This is for the safety factor.

[0065] Vertical force balance: W i =N i *cosα i +T i *sinα i Where: N i T represents the effective normal force at the bottom of the soil strip (kN / m). i The actual tangential force at the bottom of the soil strip (kN / m); Relationship between sliding force and tangential force: Actual tangential force T i Determined by the sliding force, the anti-slip force T res,i The safety factor F is provided by the shear strength of the soil. S It is the ratio of total anti-skid force to total tangential force.

[0066] The safety factor calculation submodule is used to calculate the safety factor based on the anti-sliding force and the tangential force of each soil strip.

[0067] Specifically, in the safety factor calculation submodule, the formulas for vertical equilibrium and anti-slip force are combined, and N is eliminated. i The safety factor is obtained by iterative solution, and the iterative formula is as follows:

[0068] In the formula, c i φ: Cohesion (kPa) of the soil layer containing the i-th soil strip; i : The internal friction angle (°) of the soil layer containing the i-th soil strip; W i : Weight of the i-th soil strip (kN / m); α i : The angle (°) between the bottom slip surface of the i-th soil strip and the horizontal direction; l i : Length of the slip surface at the bottom of the i-th soil strip (m).

[0069] Specifically, the iterative solution process is as follows: Initial assumption F S0 =1.0, F S0 Substituting into the right side of the formula, we obtain the first iteration value F. S1 If |F S1 -F S0 If |≤0.01, then F S1 Let F be the final safety factor; otherwise, let F S0 =F S1 Repeat the iteration until convergence; if there are multiple layers of soil and rock, substitute c according to the soil layer where the soil strip is located. i φ i Calculate and ensure parameter matching. Based on the converged F... S Sample labels are assigned according to warning level: Unstable: F S <1.00; Understability: 1.00≤F S <1.05; Basically stable: 1.05≤F S <1.25; Stable: F S ≥1.25.

[0070] Specifically, the prediction module includes: The first prediction submodule is used to acquire real-time sensing data and input the real-time sensing data into the monitoring model to obtain prediction results; The stress data acquisition submodule is used to acquire stress sensor data and ultimate bending moment of the mine support structure. Specifically, the mine support structure is generally an anti-slide pile. The stress sensor data is obtained by fiber optic grating stress sensors installed at key sections of the anti-slide pile (e.g., near the slip surface, at the top of the pile, at the bottom of the pile, or at 1 / 3 of the pile). The stress sensor data includes the normal stress and shear stress of the pile section, which are used to reflect the magnitude of the stress on the section and determine whether it is close to the ultimate stress. Fiber optic grating strain sensors are deployed every 2-3m along the longitudinal direction of the pile, covering the area above and below the slip surface, to measure the strain distribution of the pile.

[0071] The maximum bending moment calculation submodule is used to calculate the maximum bending moment of the mine support structure based on the stress sensor data. Specifically, the formula for calculating the bending moment is as follows: M=EIε; In the formula, M: measured bending moment of the anti-slide pile (kN·m), i.e., the bending moment value of the current section; E: elastic modulus of the pile material (kPa or MPa), reflecting the material stiffness; I: moment of inertia of the pile section, which is related to the shape and size of the section; ε: stress sensor data.

[0072] Bending moment is calculated using strain gauges; horizontal displacement at the pile top and rotation angle at the pile mid-section (corresponding to the slip surface) are measured using displacement gauges to reflect overall pile displacement and determine if the support effect meets standards; pore water pressure sensors are deployed in the soil around the pile (within a 5m range above and below the slip surface) to measure pore water pressure around the pile and correct for the influence of seepage on pile stress. The maximum bending moment M is calculated. 实测 and ultimate bending moment M 极限 The key ratios are then substituted to obtain the result.

[0073] The correction submodule is used to calculate the monitoring results based on the ultimate bending moment, the maximum bending moment, and the prediction results.

[0074] Specifically, based on the ultimate bending moment, the maximum bending moment, and the predicted result F... S，预测 The formula for obtaining the monitoring results is as follows: F S,修正 =F S,预测 ×[1-k(M 实测 / M 极限 -0.5)]; In the formula, M 实测 : Maximum bending moment of the anti-slide pile measured by the sensor (kN·m); M 极限 : Design ultimate bending moment of anti-slide pile (determined by the strength of pile material and cross-sectional dimensions); k: Correction coefficient (empirical value 0.3-0.5, 0.4 for open-pit mine / tunnel-landslide system).

[0075] Specifically, the data acquisition and prediction modules utilize a Bragg grating array and distributed optical fiber to collect multi-physics field data (strain, temperature, seepage pressure, vibration) from the open-pit mine. The Brillouin divergence shift (Δv) is calculated using the Brillouin divergence shift effect and the double-pulse differential method to eliminate temperature-strain cross-sensitivity, transforming the original physical quantity change signal into a desensitized signal. B The core characteristic of this sensor is that it is simultaneously affected by the coupling of temperature (ΔT) and strain (ε). The desensitization process is essentially achieved by calibrating the sensing coefficient and using dual-pulse differential separation to desensitize the coupled Δv. B The pure strain signal (ε) is calculated separately. pure ) and pure temperature signal (ΔT) pure To eliminate cross-interference, the desensitized signal is finally obtained. In open-pit mine fiber optic sensing, the Brillouin dispersion frequency shift (Δv) is... B The coupling relationship between the fiber optic cable (in MHz) and temperature change (ΔT, in °C) and strain change (ε, dimensionless, ε = ΔL / L, where ΔL is the fiber length change and ε is the original fiber length) is: Δv B =C1*ε+C2*ΔT, where C1 is the strain sensing coefficient (unit: MHz / ε), C2 is the temperature sensing coefficient (unit: MHz / ℃), ε is the axial strain of the optical fiber (true strain, including strain caused by the deformation of the open-pit mine slope), and ΔT is the temperature change of the optical fiber (relative to the calibration reference temperature T0, usually taken as 25℃). The desensitization process mainly consists of the following steps: ① Before desensitization, C1 and C2 need to be determined through calibration experiments to ensure compatibility with the characteristics of the open-pit mine optical cable: Fixed temperature: Place the mine optical cable in a constant temperature chamber (T=T0=25℃), apply known strains (ε1=0, ε2=500με, ε3=1000με), and measure the corresponding Δv. B1 Δv B2 Δv B3 Through linear fitting, we obtain C1=Δv B / ε; Fixed strain: The optical cable is fixed in a stress-free state (ε=0), and the temperature of the constant temperature chamber is adjusted (ΔT1=0℃, ΔT2=20℃, ΔT3=40℃). The corresponding Δv is measured. B4 Δv B5 Δv B6 Linear fitting yields C2=Δ vB / ΔT. ② Dual-pulse signal acquisition: Two sets of narrow pulses with a time interval of τ=100μs (pulse width 10μs, suitable for the 10m spatial resolution requirement of open-pit mines) are transmitted to the mining optical cable through a distributed optical fiber sensing system (DAS): The first set of pulses (reference pulse): Acquire the original frequency shift Δv including coupling interference. B1 The second set of pulses (probe pulses): samples the frequency shift Δv after a delay of τ at the same fiber location. B2Differential processing: Calculate the frequency shift difference Δv between the two sets of pulses. B,diff =Δv B2 -Δv B1 ③ Establish a desensitization equation set (separating pure strain and pure temperature): Since the deformation of the open-pit mine slope is a slow-changing process (strain change period ≥ 1 minute), while the temperature change is a gradual process (period ≥ 10 minutes), within the short time interval τ = 100 μs of the double pulse, the temperature can be regarded as constant (ΔT). diff =0), only the strain signal changes slightly with the slope deformation, based on which two sets of equations are established: short time scale (double pulse difference, ΔT1=ΔT2=ΔT): Subtracting the two equations eliminates ΔT, allowing direct calculation of the pure strain change Δε. pure =ε2-ε1=△v B，diff / C1.

[0076] Based on the initial calibration reference strain ε0, the current pure strain ε is obtained. pure =ε0+△v B，diff / C1; Long time scale (based on pure strain to infer pure temperature): The calculated ε pure Substituting into the core coupling formula, we can deduce the pure temperature change: △T pure =(△v B, avg -C1*ε pure ) / C2; Where Δv B,avg =(Δv B1 +Δv B2 ) / 2, which is the average frequency shift of the two sets of pulses, improving the accuracy of temperature calculation. ④ Desensitization signal output (adapted to CNN model input): Pure strain desensitization signal: ε pure (Unit: με), directly reflects the deformation state of open-pit mine slopes (without temperature interference); pure temperature-sensitive signal: ΔT pure (Unit: °C), reflecting only ambient temperature changes (without strain interference); seepage pressure / vibration correlation: ε pure Data is fused with that from a mine pore water pressure sensor and analyzed using the empirical formula (P=k*ε). pure (where k is the seepage-strain coefficient, calibrated on-site) to obtain the desensitized seepage pressure signal; the vibration signal is obtained by acquiring the instantaneous frequency shift change value through DAS, combined with ε pure The stability judgment is based on the pure vibration signal after removing strain interference.

[0077] Specifically, the neural network in the training module consists of an industrial-grade GPU (such as NVIDIA A10) and an embedded neural network chip, equipped with a customized algorithm program. The module has a dynamic weight update interface and can access the open-pit mine historical case library (such as landslide accident data) to iteratively optimize the model, which is different from the fixed model limitation of general neural network processing equipment.

[0078] Specifically, the distributed optical fiber is composed of distributed optical fiber sensing units (DAS / DTS) and equipped with MGTSV type mining flame-retardant optical cable. The module has a dustproof and impact-resistant shell (IP65 protection level), and the optical cable joint adopts special sealing parts for open-pit mines to solve the problem of equipment failure caused by dust and rainwater in open-pit mines.

[0079] Specifically, it also includes a fusion early warning module, which consists of a multi-field data fusion processor, an audible and visual early warning terminal, and a remote communication unit. The module has an emergency linkage interface that can be connected to the open-pit mine emergency system to achieve a closed loop of early warning and response, unlike traditional devices that only output early warnings.

[0080] Example 3 An example of an open-pit mine monitoring method based on fiber optic sensing from Implementation Example 1 is used. A slope engineering project in an open-pit mine is selected, and geological parameters (rock density 2300 kg / m³, elastic modulus 75 MPa, cohesion 22.5 kPa, internal friction angle 37°) and support structure parameters (anti-slide pile length 35 m, cross-section 2×2 m, spacing 4 m) are collected. A 2D model (180 m in the X direction, 70 m in the Y direction) is constructed using ABAQUS software, with free boundaries (slope surface / top) and fixed boundaries (bottom) set to simulate the stress-displacement-seepage coupling process, outputting 500 sets of simulated sensor data. The original parameters (physical parameters of soil and rock, geometric parameters of the slope, seepage environment parameters, and support structure parameters) are randomly adjusted within ±8%, generating 20,000 sets of samples, which are divided into a training set (16,000 sets) and a test set (4,000 sets) at an 8:2 ratio. A CNN model was built using Matlab software. The kernel size was 1×2, and the number of feature maps were 16, 32, 64, and 128 respectively. The learning rate was 0.001, and the model was iterated for 30 rounds. After training, the RMSE was 0.22 and the MRE was 0.08.

[0081] A fiber optic sensor array (5m spacing) was deployed on the open-pit mine slope to collect real-time strain and displacement data, which were then preprocessed and input into the model. The prediction results were corrected by combining the stress monitoring data of the anti-slide piles. The stability coefficient (and the safety factor F output by the neural network) were then considered. S When the value (which is the same as the standard value) is less than 1.05, an early warning is triggered, and support reinforcement suggestions are pushed out. The value of 1.05 is a key threshold determined by combining engineering safety, specification requirements, and practical needs, with the safety factor F... S=1.0 is the theoretical critical value (anti-sliding force = sliding force), but open-pit mines are subject to sudden disturbances such as rainstorms and vibrations, so a 5% safety margin (1.0→1.05) needs to be reserved to prevent F from being affected by sudden events. S A sharp drop to below 1.0 avoids a lack of response time under critical conditions. Referring to the "Technical Specification for Building Slope Engineering" and the "Technical Specification for Open-Pit Coal Mine Slope Engineering," the safety factor limit for the normal serviceability limit state of open-pit mine slopes is 1.05-1.10. 1.05 is a commonly used early warning threshold for high-risk slopes (such as those containing weak layers or with significant seepage). 1.05 directly relates to the risk level, engineering conditions, and specification requirements of open-pit mine slopes. Open-pit mine slopes are classified into high / medium / low risk based on height and soil stability. 1.05 corresponds to high-risk slopes (such as slope height > 40m, containing weak interlayers)—the early warning threshold for low-risk slopes can be relaxed to 1.10, while for high-risk slopes it needs to be tightened to 1.05 to match the risk level. Open-pit mine soil parameters (such as cohesion c and internal friction angle φ) have natural variability (±5%-±10%), and the threshold of 1.05 can offset the underestimation of risk caused by measured parameter errors. If the anti-slide piles and other support structures are close to their design limits (e.g., bending moment reaches 70% of the limit) and the slope's anti-slide redundancy is insufficient, the warning threshold should be lowered from 1.10 to 1.05 to trigger reinforcement in advance and prevent support structure failure. Open-pit mines are greatly affected by rainstorms, earthquakes, and mining vibrations, which can temporarily reduce F... S (For example, after a heavy rain, the pore water pressure increases, F) S (Potentially decreasing by 5%-8%), and a threshold of 1.05 provides a buffer against environmental disturbances. Comparing the predicted results with field borehole exploration data, the accuracy rate of stability level determination reached 96.3%, meeting the requirements for engineering applications.

[0082] This invention also provides a computer-readable storage medium storing a computer program thereon, which, when executed by a processor, implements any of the above-described open-pit mine monitoring methods based on fiber optic sensing.

[0083] The present invention also provides a computer device. The computer device of this invention includes: one or more processors; and a storage device for storing one or more computer programs, wherein when the one or more computer programs are executed by the one or more processors, the one or more processors implement the open-pit mine monitoring method based on fiber optic sensing provided by the present invention.

[0084] The specific embodiments described above do not constitute a limitation on the scope of protection of this invention. Those skilled in the art should understand that various modifications, combinations, sub-combinations, and substitutions can occur depending on design requirements and other factors. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of this invention should be included within the scope of protection of this invention.

Claims

1. An optical fiber sensing based monitoring method for an open pit mine, characterized by, The method comprises the following steps: S1: obtaining geological parameters and environmental parameters of a region to be monitored; S2: obtaining simulated sensing data by performing multi-physical field simulation according to the geological parameters obtained in S1; S3: obtaining expansion data by randomly expanding the environmental parameters obtained in S1; S4: obtaining a safety factor by using a limit equilibrium algorithm according to the expansion data obtained in S3; S5: inputting the simulated sensing data obtained in S2 and the safety factor obtained in S4 into a neural network model for model training to obtain a monitoring model; S6: obtaining real-time sensing data and inputting the real-time sensing data into the monitoring model obtained in S5 to obtain a monitoring result.

2. A method of monitoring an open pit mine based on fiber optic sensing as claimed in claim 1, characterized in that, In S2, the following steps are included: S21: establishing a multi-physical field 2D model according to the geological parameters obtained in S1; S22: obtaining a first potential sliding surface by using a strength reduction method according to the multi-physical field 2D model obtained in S21; S23: reading simulated sensing data of the first potential sliding surface obtained in S22 according to the multi-physical field 2D model obtained in S21.

3. A method of monitoring an open pit mine based on fiber optic sensing as claimed in claim 1, characterized in that, The geological parameters include rock mass density, elastic modulus, internal friction angle, and cohesion, and the environmental parameters include geotechnical physical parameters, slope geometric parameters, seepage environmental parameters, and supporting structure parameters.

4. A method of monitoring an open pit mine based on fiber optic sensing as claimed in claim 1, characterized in that, In S3, the expansion data is obtained by randomly adjusting the environmental parameters within a preset range.

5. A method of monitoring an open pit mine based on fiber optic sensing as claimed in claim 1, characterized in that, In S4, the following steps are included: S41: obtaining a second potential sliding surface by using a strength reduction method according to the expansion data obtained in S3; S42: dividing the second potential sliding surface obtained in S41 into a plurality of soil strips along a vertical direction; S43: calculating the length of the bottom sliding surface of each soil strip and the weight of each soil strip according to the expansion data obtained in S3; S44: calculating the sliding resistance and tangential force of each soil strip according to the stress balance analysis, the length of the bottom sliding surface of each soil strip obtained in S43, and the weight of each soil strip obtained in S43; S45: calculating a safety factor according to the sliding resistance of each soil strip obtained in S44 and the tangential force of each soil strip obtained in S44.

6. A method of monitoring an open pit mine based on fiber optic sensing as claimed in claim 1, characterized by, The monitoring model comprises: an input layer for receiving the simulated sensing data obtained in S2 and the safety factor obtained in S4; a convolution layer for obtaining local spatial correlation between the simulated sensing data and the safety factor; a time series analysis layer for predicting a stability change trend of a slope according to the simulated sensing data and the safety factor to obtain a trend vector T; a fully connected layer including two hidden layers, each of which is provided with an activation function, for dynamically adjusting the weights of the simulated sensing data and the safety factor and fusing the trend vector T; an output layer for outputting a monitoring result, wherein the monitoring result includes a current safety factor and a safety level corresponding to the current safety factor.

7. A method of monitoring an open pit mine based on fiber optic sensing as claimed in claim 1, characterized by, In S6, the following steps are included: S61: obtaining real-time sensing data and inputting the real-time sensing data into the monitoring model obtained in S5 to obtain a prediction result; S62: obtaining stress sensor data of a mine supporting structure and a limit bending moment of the mine supporting structure; S63: calculating the maximum bending moment of the mine support structure according to the stress sensor data obtained in S62; S64: calculating the monitoring result according to the limit bending moment obtained in S62, the maximum bending moment obtained in S63 and the prediction result obtained in S61.

8. An optical fiber sensing based monitoring device for an open pit mine, characterized in that, The open-pit mine monitoring method based on optical fiber sensing according to claim 1, comprising: a data acquisition module configured to acquire geological parameters and environmental parameters of a region to be monitored; a simulation module configured to perform multi-physical field simulation to obtain simulation sensing data according to the geological parameters; a capacity expansion module configured to perform random capacity expansion on the environmental parameters to obtain expanded data; a safety factor calculation module configured to calculate a safety factor according to the expanded data using a limit equilibrium algorithm; a training module configured to input the simulation sensing data and the safety factor as training samples into a neural network model to perform model training and obtain a monitoring model; a prediction module configured to acquire real-time sensing data and input the real-time sensing data into the monitoring model to obtain a monitoring result.

9. A computer device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, A processor executes a computer program to implement an open-pit mine monitoring method based on optical fiber sensing according to any one of claims 1-7.

10. A computer-readable storage medium having stored thereon a computer program, characterized in that, The computer program is executed by the processor to implement an open-pit mine monitoring method based on optical fiber sensing according to any one of claims 1-7.

Images