Thin-layer weakly cemented expansive rock tunnel safety risk grading method and supporting structure

Abstract

The invention provides a thin-layer weakly-cemented expansive rock tunnel safety risk grading method and a supporting structure, and relates to the technical field of tunnel surrounding rock engineering. Carrying out stratification feature recognition and index quantification on the expansive rock geological parameters, and constructing an index comprehensive judgment model; inputting the index comprehensive judgment model into a fuzzy neural network algorithm for coupling processing, and establishing a nonlinear risk evaluation model; according to a reliability theory and Monte Carlo simulation, determining a quantitative relation between a risk index and a surrounding rock instability probability in the nonlinear risk evaluation model, and constructing a dynamic risk threshold value; and performing grade division on tunnel construction risks based on the dynamic risk threshold, and performing dynamic feedback and parameter updating on a preset risk evaluation model in combination with real-time monitoring data to obtain a tunnel safety risk grading scheme. According to the invention, the prominent problem of unbalanced engineering safety risk and cost control is solved.

Description

Technical Field

[0001] This invention relates to the field of tunnel surrounding rock engineering technology, and more specifically, to a method for safety risk classification and support structure for thin-layer weakly cemented expansive rock tunnels. Background Technology

[0002] In existing tunnel surrounding rock engineering technologies, risk classification and support design mainly rely on the BQ method in the "Highway Tunnel Design Code" or the RMR and Q system methods used abroad. These traditional methods are based on a two-dimensional evaluation framework of rock hardness and rock mass integrity, focusing on the uniaxial compressive strength of the rock and the rock mass integrity index. They are insufficiently applicable to thin-layered, weakly cemented rock strata with significant expansibility. Because such rock masses have low strength, short self-stabilization time, are easily softened and expanded upon contact with water, and their deformation is controlled by structural surfaces, traditional methods are unable to effectively predict the risk of large deformations. This results in a lack of reliable basis for support design, leading to an over-reliance on engineering experience. This reliance on experience leads to two problems: first, overly conservative designs, resulting in wasted materials and costs; second, insufficient support strength, failing to suppress the development of surrounding rock deformation, and consequently causing large tunnel deformation, support structure failure, or even collapse and other safety accidents. Ultimately, this results in a prominent imbalance between engineering safety risks and cost control.

[0003] Therefore, there is an urgent need for a safety risk classification method and support structure for thin-layer weakly cemented expansive rock tunnels, which solves the prominent problem of imbalance between engineering safety risk and cost control. Summary of the Invention

[0004] The purpose of this invention is to provide a safety risk classification method and support structure for tunnels in thin-layer, weakly cemented, expansive rock, in order to improve the aforementioned problems. To achieve the above objective, the technical solution adopted by this invention is as follows:

[0005] In a first aspect, this application provides a method for safety risk classification of tunnels in thin-layered, weakly cemented, expansive rock, characterized by comprising:

[0006] Obtain geological parameters of the expansive rock;

[0007] The geological parameters of the expansive rock are subjected to bedding feature identification and index quantification to construct a comprehensive index judgment model;

[0008] The comprehensive judgment model of the indicators is input into the fuzzy neural network algorithm for coupling processing to establish a nonlinear risk assessment model;

[0009] Based on reliability theory and Monte Carlo simulation, the quantitative relationship between the risk index and the probability of surrounding rock instability in the nonlinear risk assessment model is determined, and a dynamic risk threshold is constructed.

[0010] Based on the dynamic risk threshold, the tunnel construction risk is classified into levels, and the preset risk assessment model is dynamically fed back and its parameters are updated in combination with real-time monitoring data to obtain a tunnel safety risk classification scheme.

[0011] Secondly, this application also provides a safety risk support structure for thin-layer, weakly cemented, expansive rock tunnels, including:

[0012] Different risk sections in the tunnel safety risk classification scheme are configured with graded and adapted support systems. The risk sections correspond to conventional, reinforced, high-strength, and composite retractable support structures in order from lightest to highest risk.

[0013] The beneficial effects of this invention are as follows:

[0014] This invention constructs a comprehensive index-based judgment model, transforming qualitative bedding characteristics into quantitative evaluation indicators to reflect the engineering characteristics of real thin-layered, weakly cemented rock masses. Based on this, it utilizes the self-learning and nonlinear mapping capabilities of fuzzy neural networks, combined with the advantages of fuzzy theory in handling uncertain information, to improve the accuracy of identifying surrounding rock instability risks under complex geological conditions and reduce the false judgment rate. Furthermore, based on reliability theory and Monte Carlo simulation, a quantitative relationship between the risk index and the probability of surrounding rock instability is established to adapt to the differentiated risk management needs under different geological conditions. Combined with real-time monitoring data, the risk assessment model is dynamically fed back and its parameters are updated, allowing the risk classification scheme to be adjusted in real time according to the construction progress. This enables early warning of potential disasters, thereby reducing the probability of construction safety accidents. In summary, this invention solves the prominent problem of the imbalance between engineering safety risk and cost control.

[0015] Other features and advantages of the invention will be set forth in the following description, and will be apparent in part from the description, or may be learned by practicing embodiments of the invention. The objects and other advantages of the invention may be realized and obtained by means of the structures particularly pointed out in the written description, claims, and drawings. Attached Figure Description

[0016] To more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings used in the embodiments will be briefly introduced below. It should be understood that the following drawings only show some embodiments of the present invention and should not be regarded as a limitation on the scope. For those skilled in the art, other related drawings can be obtained based on these drawings without creative effort.

[0017] Figure 1 This is a schematic diagram of the safety risk classification method for thin-layer weakly cemented expansive rock tunnels as described in this embodiment of the invention;

[0018] Figure 2This is a schematic diagram of the safety risk classification device for thin-layer weakly cemented expansive rock tunnels as described in an embodiment of the present invention.

[0019] The markings in the diagram are: 800, Safety Risk Classification Equipment for Thin-Layer Weakly Cemented Expanded Rock Tunnels; 801, Processor; 802, Memory; 803, Multimedia Components; 804, I / O Interface; 805, Communication Components. Detailed Implementation

[0020] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. The components of the embodiments of the present invention described and shown in the accompanying drawings can generally be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of the present invention provided in the accompanying drawings is not intended to limit the scope of the claimed invention, but merely to illustrate selected embodiments of the invention. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without inventive effort are within the scope of protection of the present invention.

[0021] It should be noted that similar reference numerals and letters in the following figures indicate similar items; therefore, once an item is defined in one figure, it does not need to be further defined and explained in subsequent figures. Furthermore, in the description of this invention, terms such as "first," "second," etc., are used only to distinguish descriptions and should not be construed as indicating or implying relative importance.

[0022] In existing tunnel surrounding rock engineering technologies, risk classification and support design are mainly based on a two-dimensional evaluation framework centered on rock hardness and rock mass integrity. However, for thin-layered, weakly cemented expansive rock strata, on the one hand, the uniaxial saturated compressive strength test of the rock, on which it relies, is difficult to accurately reflect the true strength characteristics of the weakly cemented rock mass; on the other hand, the existing evaluation system does not include key hydrological parameters such as expansibility and disintegration in quantitative indicators, resulting in the complete neglect of the mechanical deterioration mechanism of the rock mass after encountering water, thus causing a serious disconnect between the surrounding rock evaluation results in the exploration stage and the actual risks during construction. In addition, for thin-layered rock masses with bedding spacing of less than 0.3m, their structural surface density can reach 5-10 times that of traditional blocky rock masses, while the current specifications only provide a qualitative description through "joint development degree," which cannot quantify the control effect of bedding occurrence on the stability of the surrounding rock, further weakening the pertinence and reliability of classification and support design.

[0023] Example 1:

[0024] This embodiment provides a method for safety risk classification of tunnels in thin-layered, weakly cemented, expansive rock.

[0025] See Figure 1 The figure shows that the method includes steps S1 to S5, including:

[0026] S1: Obtain geological parameters of the swelling rock;

[0027] S2: Identify the bedding features and quantify the indexes of the geological parameters of the expansive rock, and construct a comprehensive index judgment model;

[0028] To clarify the specific method for obtaining the comprehensive indicator judgment model, step S2 includes S21 to S24, specifically:

[0029] S21: Perform multiple characterization processes on the layer thickness index in the geological parameters of the swelling rock to obtain quantitative layer thickness classification values;

[0030] In this step, the layer thickness index in the geological parameters of the expansive rock is subjected to multiple characterization processes including digitalization, automation, and probabilistic analysis. The bedding planes are automatically identified and the layer thickness is calculated using an image recognition algorithm. The layer thickness is then classified into four levels based on the probability of occurrence: extremely thin layer, thin layer, medium-thick layer, and thick layer, to obtain a quantitative layer thickness classification value.

[0031] The extremely thin layer (D0) is: layer thickness < 0.1m, assigned a value of 4;

[0032] The thin layer (D1) is defined as having a thickness of 0.1m ≤ layer thickness < 0.3m and is assigned a value of 3.

[0033] The medium-thickness layer (D2) is defined as: 0.3m ≤ layer thickness < 0.7m, assigned a value of 2;

[0034] The thick layer (D3) is defined as having a thickness ≥ 0.7m and a value of 1.

[0035] S22: Based on preset main control indicators and verification indicators, the rock mass structure characteristic indicators in the geological parameters of the expansive rock are comprehensively judged to obtain the quantitative value of the rock mass structure.

[0036] In this step, the micro-drilling resistance test results are used as the main control index, and the disintegration test data and mineral composition analysis results are used as verification indexes. Weighted calculations determine the cementation level as weak, moderate, or strong cementation, and a quantitative value for the degree of cementation is output. Further, expansion strain tests are conducted on cubic samples along the bedding normal and parallel directions. Combining the expansion force-time curve characteristics and the joint analysis of montmorillonite content-specific surface area, the expansion level is comprehensively determined as strong, moderate, weak, or slight expansion, and a quantitative value for expansion is output. Then, based on borehole, adit, and surface outcrop data, a three-dimensional attitude field model is constructed using Kriging interpolation. Attitude parameters are extracted along the tunnel axis, and geometric correction coefficients are introduced based on the tunnel span and burial depth to calculate a quantitative value for the attitude effect reflecting spatial variability. Finally, the quantitative values ​​for the degree of cementation, expansion, and attitude effect are integrated to obtain a quantitative value for the rock mass structure.

[0037] S23: Based on the confidence level, the quantitative layer thickness classification value and the quantitative rock mass structure assignment are evaluated and the extreme value data of the index are screened. The extreme value data of the index are dynamically updated by combining the range standardization to obtain the standardized index value.

[0038] In this step, the quantitative layer thickness classification value and the quantitative rock mass structure assignment value are divided into Grade A, Grade B and Grade C according to the confidence level, and quality evaluation and screening are carried out. Grade A data is used directly, Grade B data is increased by 20% safety factor, and Grade C data is only used for auxiliary verification. After extracting the extreme value data of the index, the extreme value data of the index is dynamically updated by combining the range standardization method and the standardization benchmark is continuously optimized to obtain the standardized index value.

[0039] S24: Based on the multiple linear regression algorithm, the standardized index values ​​and the preset index weight system are constructed to obtain the comprehensive index judgment model.

[0040] This step considers the nonlinear coupling effect between indicators such as layer thickness, degree of cementation, expansibility, and rock stratum attitude. It uses a multiple linear regression algorithm combined with interaction term correction to construct the standardized indicator values ​​and the preset indicator weight system, thereby obtaining an indicator comprehensive judgment model. The indicator comprehensive judgment model is used for safety risk prediction of expansive rock tunnels and for construction risk assessment in engineering scenarios such as high-speed railway tunnels and highway tunnels that pass through thin, weakly cemented expansive rock strata.

[0041] S3: Input the comprehensive judgment model of the indicators into the fuzzy neural network algorithm for coupling processing to establish a nonlinear risk assessment model;

[0042] To clarify the specific method for obtaining the nonlinear risk assessment model, step S3 includes S31 to S34, specifically:

[0043] S31: Conduct orthogonal numerical tests on the comprehensive judgment model of the index, combine it with the three-dimensional numerical simulation model of geotechnical engineering to conduct surrounding rock response, and construct a nonlinear coupling effect model.

[0044] In this step, the geotechnical engineering three-dimensional numerical simulation model is the geotechnical engineering three-dimensional numerical simulation software (FLAC3D).

[0045] To clarify the specific method for obtaining the nonlinear coupling effect model, step S31 includes S311 to S315, specifically:

[0046] S311: The indicators affecting the stability of the surrounding rock in the comprehensive judgment model are used to design multiple sets of numerical simulation working conditions with different indicator levels through orthogonal experiments to obtain a combination of mechanical response indicators.

[0047] In this step, the core indicators affecting the stability of the surrounding rock—layer thickness, degree of cementation, expansibility, and stratum attitude—are extracted from the comprehensive index judgment model. Key influencing indicators are screened using the Plackett-Burman experimental design method. Then, orthogonal experimental design is employed to design 32 sets of numerical simulation conditions with different levels of key indicators to simulate the stability of the surrounding rock, resulting in a combination of mechanical response indicators that includes surrounding rock displacement and the development of the plastic zone. This step simplifies the indicator dimensions through Plackett-Burman design and leverages the balance of orthogonal experiments to ensure the representativeness of the combined working conditions and fully cover the coupling effects between key indicators.

[0048] S312: Perform geotechnical engineering three-dimensional numerical simulation model calculations on the combination of mechanical response indicators to obtain surrounding rock stability evaluation indicators;

[0049] In this step, the specific calculations of the three-dimensional numerical simulation model are as follows: the model parameters are calibrated in conjunction with actual engineering geological conditions to ensure the reliability and engineering applicability of the simulation results. Subsequently, by reasonably setting boundary conditions, the entire process of stress and deformation of the surrounding rock under different combinations of mechanical response indices is simulated, and finally, the stability evaluation index of the surrounding rock is obtained.

[0050] The surrounding rock stability evaluation indicators include surrounding rock convergence displacement, plastic zone depth, and surrounding rock pressure.

[0051] S313: The relationship between the combination of mechanical response indices and the surrounding rock stability evaluation indices is fitted and analyzed using the response surface methodology to obtain a nonlinear model;

[0052] In this step, the relationship between the combination of mechanical response indicators and the evaluation indicators of surrounding rock stability is fitted and analyzed using the response surface methodology. During the process, the model parameters are optimized using the least squares method to determine the weights of the indicators in the combination of mechanical response indicators. The main effects and interaction effects of the evaluation indicators of surrounding rock stability are analyzed using the mechanical response indicators. For the two key sets of indicators, thin-layer structure and expansibility, and cementation degree and rock stratum attitude, fitting coupling terms are used to characterize the nonlinear interaction between them, resulting in a nonlinear model.

[0053] The nonlinear model is a nonlinear model that includes first-order terms, second-order terms, and interaction terms.

[0054] S314: Based on regression analysis, establish the relationship between the surrounding rock stability evaluation index and the surrounding rock response, and construct a regression coefficient matrix;

[0055] In this step, a quantitative relationship between the surrounding rock stability evaluation index and the surrounding rock response is established based on regression analysis. By statistically analyzing the numerical simulation results, the first-order effect coefficient, second-order nonlinear coefficient, and interaction coefficient corresponding to each index in the nonlinear model are solved, and a regression coefficient matrix is ​​constructed.

[0056] S315: Using the preset dimensionless displacement as the risk response value, the nonlinear model is corrected by substituting the risk response value into the regression coefficient matrix to obtain the nonlinear coupling effect model.

[0057] In this step, the dimensionless displacement is calculated by the ratio of the actual displacement to the tunnel span. The nonlinear model is corrected by substituting the risk response value into the regression coefficient matrix. The coefficients of the nonlinear model are adjusted to match the correspondence between the risk response value and the index combination, and finally a nonlinear coupling effect model is obtained. The nonlinear coupling effect model is used to reflect the index coupling effect.

[0058] S32: Input the geological parameters of the swelling rock into a fuzzy neural network algorithm for adaptive neural fuzzy inference training to construct a risk assessment model;

[0059] To clarify the specific method for obtaining the risk assessment model, step S32 includes S321 to S324, which are as follows:

[0060] S321: Using the measured data in the geological parameters of the expansive rock as the basic sample, and combining it with the expanded data in the three-dimensional numerical simulation model of geotechnical engineering and the expert experience sample data, a hybrid sample library is constructed.

[0061] In this step, the measured data of the geological parameters of the expansive rock are used as the basic sample. Combined with the parameters in the three-dimensional numerical simulation model of geotechnical engineering (FLAC3D) to generate 5,000 sets of extended data through full-coverage simulation, and 300 sets of experience sample data obtained by inviting 15 senior experts to score, a total of 7,650 sets of mixed sample library are constructed. The mixed sample library is divided into training set, validation set and test set in a ratio of 7:2:1.

[0062] The measured data includes 2,350 sets of historical engineering case data from 47 expansive rock tunnel surrounding rock projects, as well as measured values ​​of core indicators such as layer thickness, cementation degree, expansibility, and rock stratum occurrence.

[0063] S322: Input the hybrid sample library into the adaptive neurofuzzy inference model for network training and optimization to obtain the adaptive neurofuzzy risk assessment model;

[0064] In this step, the hybrid sample library is input into the Adaptive Neural Fuzzy Inference Model (ANFIS) for network training and optimization. The network structure is set as follows: input layer with 4 nodes (corresponding to standardized layer thickness, cementation degree, dilatancy, and rock stratum attitude); fuzzification layer with each index divided into 3 fuzzy sets (low, medium, and high) using Gaussian membership functions; rule layer with 81 rules (corresponding to different risk modes); and defuzzification layer outputting the accurate risk index. The training algorithm uses a hybrid training method of LM optimization algorithm and gradient descent method, and the training error is reduced by iteratively adjusting the model parameters. Training is stopped when the training error MSE ≤ 0.032 and the validation error MSE ≤ 0.041, and the adaptive neural fuzzy risk assessment model is output.

[0065] S323: Based on the adaptive neurofuzzy risk assessment model, calculate the comprehensive risk index and perform sensitivity analysis on the target tunnel to obtain the comprehensive risk index and sensitivity coefficient;

[0066] S324: Adjust the feedback weights of the adaptive neural fuzzy inference model based on the comprehensive risk index and the sensitivity coefficient to obtain the risk assessment model.

[0067] In this step, the adaptive neurofuzzy inference model is adjusted by feedback weights based on the comprehensive risk index and the sensitivity coefficient. The model is optimized by combining the matching degree between the comprehensive risk index and the actual engineering risk and the rationality of the sensitivity coefficient. The membership function parameters and rule layer weights are dynamically adjusted through the backpropagation mechanism so that the adaptive neurofuzzy inference model reflects the nonlinear relationship between indicators, and finally the risk assessment model is obtained.

[0068] S33: Based on Monte Carlo simulation, large-scale random sampling calculations are performed on the risk assessment model and the nonlinear coupling effect model to obtain the predicted probability distribution and confidence interval of the risk index;

[0069] In this step, a large-scale random sampling calculation is performed on the risk assessment model and the nonlinear coupling effect model based on Monte Carlo simulation. The number of samplings is set to 10,000. Considering the measurement error of the input parameters and the uncertainty of the model structure, the risk index corresponding to each sampling is output, and the predicted probability distribution and 95% confidence interval of the risk index are statistically obtained.

[0070] S34: The confidence interval is compared with the preset confidence constraints. When the confidence constraints are met, the risk assessment model corresponding to the predicted probability distribution of the risk index is used as the nonlinear risk assessment model.

[0071] In this step, the confidence interval is compared with the preset confidence constraint (confidence interval width). The system makes a judgment: when the confidence constraint is met, it indicates that the model uncertainty is within an acceptable range, and the risk assessment model corresponding to the predicted probability distribution of the risk index is used as a nonlinear risk assessment model; when the confidence interval width ΔR > 0.5, a data supplementation and collection procedure is triggered to increase the number of test samples and reduce parameter uncertainty. The model is then retrained and Monte Carlo simulation is performed until the confidence constraint is met, and the risk assessment model corresponding to the predicted probability distribution of the risk index is used as a nonlinear risk assessment model.

[0072] S4: Based on the reliability theory and Monte Carlo simulation, the quantitative relationship between the risk index and the probability of surrounding rock instability in the nonlinear risk assessment model is determined, and a dynamic risk threshold is constructed.

[0073] To clarify the specific method for obtaining the dynamic risk threshold, step S4 includes S41 to S43, specifically:

[0074] S41: Based on reliability theory and Monte Carlo simulation, the mechanical response of surrounding rock under different risk indices is stochastically simulated and statistically analyzed using a three-dimensional numerical simulation model for geotechnical engineering to obtain the instability probability of the plastic zone.

[0075] In this step, based on reliability theory and Monte Carlo simulation, the mechanical response of surrounding rock under different risk indices is randomly simulated and statistically analyzed using the three-dimensional numerical simulation model of geotechnical engineering (FLAC3D). The number of simulations is set to 125 groups. Considering the parameter variation characteristics, the convergence displacement ≥ 0.1 tunnel span is used as the criterion for surrounding rock instability. The instability situation corresponding to different risk indices is statistically analyzed to obtain the instability probability of the plastic zone.

[0076] S42: Based on the quantitative relationship between the risk index and the probability of instability in the plastic zone fitted by logistic regression, the initial risk level is divided by combining the reliability index of the engineering structure to obtain the risk level standard;

[0077] In this step, the quantitative relationship between the risk index and the instability probability of the plastic zone is fitted based on logistic regression. The target reliability index is determined in combination with the "Unified Standard for Reliability Design of Engineering Structures" and the characteristics of the expansive rock strata. The reliability index is then mapped back to the risk index domain through the instability probability to divide the initial risk level threshold. Finally, a risk level standard including Level I (very low) to Level V (very high) is constructed.

[0078] The aforementioned risk level criteria:

[0079] Level I (Extremely Low Risk): Risk Index R < 1.8 (Probability of Surrounding Rock Instability Pf < 10) -5 (Reliability index β > 4.3).

[0080] Level II (Low Risk): 1.8 ≤ Risk Index R < 2.6 (10 -5 ≤Probability of surrounding rock instability Pf<5×10 -4 (3.5 < reliability index β ≤ 4.3).

[0081] Level III (Medium Risk): 2.6 ≤ Risk Index R < 3.4 (5 × 10⁻⁶) -4 ≤Probability of surrounding rock instability Pf<2×10 -2 (2.9 < reliability index β ≤ 3.5).

[0082] Level IV (High Risk): 3.4 ≤ Risk Index R < 4.2 (2 × 10⁻⁶) -2 ≤Probability of surrounding rock instability Pf<5×10 -2 (2.6 < reliability index β ≤ 2.9).

[0083] Level V (Extremely High Risk): Risk index R ≥ 4.2 (Probability of surrounding rock instability Pf ≥ 5 × 10⁻⁶) -2 (Reliability index β≤2.6)

[0084] S43: Based on the dynamic adjustment mechanism, the risk level standard is corrected in real time by monitoring displacement rate, groundwater flow and blasting vibration velocity, and a dynamic risk threshold is constructed by combining the time evolution characteristics of the risk index.

[0085] In this step, the real-time monitoring data is used to correct the risk level standards. The risk level is automatically increased by one level when the monitored displacement rate is greater than 5 mm / d or the groundwater flow rate is greater than 10 L / min·m, and by half a level when the blasting vibration velocity is greater than 15 cm / s. Simultaneously, a time evolution equation for the risk index is established, and a dynamic update frequency is set according to the tunnel construction progress: daily updates when the excavation face is 0.5B from the monitoring section, updates every 3 days when it is 1B-2B, and weekly updates when it is greater than 2B. Finally, evaluation units are divided every 5m along the tunnel axis, and spatial refinement of the risk level is achieved through Kriging interpolation. The time evolution equation of the risk index is integrated with the spatial division results to generate a dynamic risk threshold, which adapts in real-time to the spatiotemporal evolution of the surrounding rock conditions during construction.

[0086] The time evolution equation of the risk index is expressed as follows:

[0087] (1)

[0088] In the above formula (1), For a moment The corresponding real-time risk index, To initially evaluate the risk index, It is a natural constant. The time degradation factor, For time variables, This is a comprehensive monitoring and correction item.

[0089] S5: Based on the dynamic risk threshold, the tunnel construction risk is classified into levels, and the preset risk assessment model is dynamically fed back and its parameters are updated in combination with real-time monitoring data to obtain a tunnel safety risk classification scheme.

[0090] To clarify the specific method for obtaining the tunnel safety risk classification scheme, step S5 includes S51 to S53, specifically:

[0091] S51: Based on the dynamic risk threshold and model self-learning mechanism, the error between the risk prediction value and the measured deformation is calculated by real-time monitoring data to obtain the error value;

[0092] In this step, based on the dynamic risk threshold and model self-learning mechanism, the deviation rates of key indicators such as displacement, anchor bolt axial force and surrounding rock pressure relative to the model prediction values ​​are first calculated according to preset frequencies using real-time monitoring data collected by multi-source devices such as total station, laser scanning, distributed optical fiber and flexible sensor array. The error values ​​are then obtained by weighted algorithm and used to reflect the difference between the model prediction and the actual working conditions.

[0093] S52: When the error value exceeds the preset threshold, the incremental learning algorithm is triggered to train and update the parameters of the preset risk assessment model using real-time geological information and the real-time monitoring data to obtain an updated risk assessment model.

[0094] In this step, when the error value When the cumulative number of valid samples exceeds 5, the incremental learning algorithm is automatically triggered. This algorithm uses real-time geological information from the working face and the real-time monitoring data as new training samples to iteratively train and update the parameters of the pre-defined risk assessment model. During the update process, while retaining the model's historical knowledge, the algorithm focuses on optimizing its membership function parameters, rule layer weights, and regression coefficient matrix. To ensure update quality, the entire incremental learning process must meet strict performance control standards: the mean squared error (MSE) of training must not exceed 0.041, and the prediction accuracy of the updated model on the test set (within ±10%) must not be less than 89.3%, effectively avoiding overfitting. The updated risk assessment model is then obtained.

[0095] S53: Based on the updated risk assessment model and the dynamic risk threshold, the evaluation units along the tunnel are dynamically classified into risk levels to construct a tunnel safety risk classification scheme.

[0096] In this step, based on the updated risk assessment model and dynamic risk threshold, a preliminary risk assessment is first conducted along the tunnel axis, with each assessment unit measuring 5 meters. A continuous risk index cloud map is then generated using Kriging interpolation, achieving spatially refined risk level classification and accurate identification of high-risk "hotspot" sections. Based on this, the risk level is periodically and dynamically updated using the time evolution equation of the risk index. The update frequency is set according to the construction progress (daily updates when the excavation face is 0.5B from the monitoring section, every 3 days when it is 1B-2B, and weekly updates when it is >2B). The classification process strictly follows the five-level standard based on reliability theory, and the monitoring displacement rate is also considered. Groundwater flow or blast vibration speed The system automatically triggers a risk level adjustment mechanism, ultimately resulting in a tunnel safety risk classification scheme.

[0097] Example 2:

[0098] This embodiment provides a safety risk support structure for thin-layer weakly cemented expansive rock tunnels, the structure comprising:

[0099] Different risk sections in the tunnel safety risk classification scheme are configured with graded and adapted support systems. The risk sections correspond to conventional, reinforced, high-strength, and composite retractable support structures in order from lightest to highest risk.

[0100] Conventional support structure: The first and second risk sections among the multiple risk sections are supported by ordinary advanced anchor bolts, I-beam arch frames, system mortar anchor bolts and C25 shotcrete.

[0101] Reinforced support structure: For the third-level risk section among the multiple risk sections, a combination of encrypted advanced small pipe grouting support, I-beam arch frame, extended system anchor bolts, steel mesh, locking anchor pipes, and C25 shotcrete is used for support.

[0102] High-strength support structure: For the fourth-level risk section among the multiple risk sections, pre-grouting is used for pre-support, and H-shaped steel arch frame, heavy I-beam steel arch frame, hollow grouting anchor rod, double-layer steel mesh and C25 shotcrete are configured. Strong locking foot structure and invert arch must be set to form a high-strength support structure.

[0103] Composite retractable support structure: The fifth-level risk section among the multiple risk sections is supported by a strong support structure, retractable joints and adaptive support modules to form a composite retractable support.

[0104] The retractable joint is equipped with a specially designed U-shaped cable clamp or hydraulic retractable component at the central connection of the invert arch, which allows the support structure to undergo limited and controllable displacement under expansion stress in order to release some energy and avoid stress accumulation causing brittle failure.

[0105] The adaptive support module embeds pressure sensors and monitoring units within the support system to monitor surrounding rock pressure and support deformation in real time, and is linked to the grouting system. When the pressure exceeds a threshold, supplementary grouting is automatically initiated to enhance the self-supporting capacity of the surrounding rock and achieve dynamic optimization of the support-surrounding rock interaction.

[0106] It should be noted that the specific methods by which each module performs its operations in the above embodiments have been described in detail in the embodiments of the method, and will not be elaborated here.

[0107] Example 3:

[0108] Corresponding to the above method embodiments, this embodiment also provides a safety risk classification device for thin-layer weakly cemented expansive rock tunnels. The safety risk classification device for thin-layer weakly cemented expansive rock tunnels described below can be referred to in correspondence with the safety risk classification method for thin-layer weakly cemented expansive rock tunnels described above.

[0109] Figure 2 This is a block diagram illustrating a safety risk classification device 800 for a thin-layer, weakly cemented, expansive rock tunnel according to an exemplary embodiment. Figure 2As shown, the thin-layer weakly cemented expansive rock tunnel safety risk classification device 800 may include: a processor 801 and a memory 802. The thin-layer weakly cemented expansive rock tunnel safety risk classification device 800 may also include one or more of the following: a multimedia component 803, an I / O interface 804, and a communication component 805.

[0110] The processor 801 controls the overall operation of the thin-layer weakly cemented expansive rock tunnel safety risk classification device 800 to complete all or part of the steps in the aforementioned thin-layer weakly cemented expansive rock tunnel safety risk classification method. The memory 802 stores various types of data to support the operation of the thin-layer weakly cemented expansive rock tunnel safety risk classification device 800. This data may include, for example, instructions for any application or method operating on the thin-layer weakly cemented expansive rock tunnel safety risk classification device 800, as well as application-related data, such as contact data, sent and received messages, images, audio, video, etc. The memory 802 can be implemented using any type of volatile or non-volatile storage device or a combination thereof, such as Static Random Access Memory (SRAM), Electrically Erasable Programmable Read-Only Memory (EEPROM), Erasable Programmable Read-Only Memory (EPROM), Programmable Read-Only Memory (PROM), Read-Only Memory (ROM), magnetic storage, flash memory, magnetic disk, or optical disk. The multimedia component 803 may include a screen and an audio component. The screen may be, for example, a touchscreen, and the audio component is used to output and / or input audio signals. For example, the audio component may include a microphone for receiving external audio signals. The received audio signals may be further stored in the memory 802 or transmitted via the communication component 805. The audio component also includes at least one speaker for outputting audio signals. I / O interface 804 provides an interface between processor 801 and other interface modules, such as keyboards, mice, and buttons. These buttons can be virtual or physical. Communication component 805 is used for wired or wireless communication between the thin-layer weakly cemented expansive rock tunnel safety risk classification device 800 and other devices. Wireless communication includes Wi-Fi, Bluetooth, Near Field Communication (NFC), 2G, 3G, or 4G, or a combination thereof. Therefore, the corresponding communication component 805 may include a Wi-Fi module, a Bluetooth module, or an NFC module.

[0111] In an exemplary embodiment, the thin-layer weakly cemented expansive rock tunnel safety risk classification device 800 may be implemented by one or more application-specific integrated circuits (ASICs), digital signal processors (DSPs), digital signal processing devices (DSPDs), programmable logic devices (PLDs), field programmable gate arrays (FPGAs), controllers, microcontrollers, microprocessors, or other electronic components to perform the above-described thin-layer weakly cemented expansive rock tunnel safety risk classification method.

[0112] Example 4:

[0113] Corresponding to the above method embodiments, this embodiment also provides a medium. The medium described below can be referred to in conjunction with the above-described method for safety risk classification of thin-layer weakly cemented expansive rock tunnels.

[0114] A medium storing a computer program, which, when executed by a processor, implements the steps of the method for safety risk classification of thin-layer weakly cemented expansive rock tunnels as described in the above method embodiments.

[0115] The medium can specifically be any medium capable of storing program code, such as a USB flash drive, external hard drive, read-only memory (ROM), random access memory (RAM), magnetic disk, or optical disk.

[0116] The above description is merely a preferred embodiment of the present invention and is not intended to limit the invention. Various modifications and variations can be made to the present invention by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

[0117] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the technical scope disclosed in the present invention should be included within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the scope of the claims.

Claims

1. A method for safety risk classification of tunnels in thin-layered, weakly cemented expansive rock, characterized in that, include: Obtain geological parameters of the expansive rock; The geological parameters of the expansive rock are subjected to bedding feature identification and index quantification to construct a comprehensive index judgment model; The comprehensive judgment model of the indicators is input into the fuzzy neural network algorithm for coupling processing to establish a nonlinear risk assessment model; Based on reliability theory and Monte Carlo simulation, the quantitative relationship between the risk index and the probability of surrounding rock instability in the nonlinear risk assessment model is determined, and a dynamic risk threshold is constructed. Based on the dynamic risk threshold, the tunnel construction risk is classified into levels, and the preset risk assessment model is dynamically fed back and its parameters are updated in combination with real-time monitoring data to obtain a tunnel safety risk classification scheme.

2. The method for safety risk classification of thin-layer weakly cemented expansive rock tunnels according to claim 1, characterized in that, The geological parameters of the expansive rock are subjected to bedding feature identification and index quantification, and a comprehensive index judgment model is constructed, including: Multiple characterization processes were performed on the layer thickness index in the geological parameters of the expansive rock to obtain quantitative layer thickness classification values. Based on preset main control indicators and verification indicators, the rock mass structure characteristic indicators in the geological parameters of the expansive rock are comprehensively judged to obtain the quantitative value of the rock mass structure. Based on the confidence level, the quantitative layer thickness classification value and the quantitative rock mass structure value are evaluated and the extreme value data of the index are screened. The extreme value data of the index are dynamically updated by combining the range standardization to obtain the standardized index value. Based on the multiple linear regression algorithm, the standardized index values ​​and the preset index weight system are constructed to obtain the comprehensive index judgment model.

3. The method for safety risk classification of thin-layer weakly cemented expansive rock tunnels according to claim 1, characterized in that, The comprehensive judgment model based on the aforementioned indicators is input into a fuzzy neural network algorithm for coupling processing to establish a nonlinear risk assessment model, including: The comprehensive judgment model of the index is subjected to orthogonal numerical tests, and the surrounding rock response is combined with the three-dimensional numerical simulation model of geotechnical engineering to construct a nonlinear coupling effect model. The geological parameters of the swelling rock are input into a fuzzy neural network algorithm for adaptive neural fuzzy inference training to construct a risk assessment model. Based on Monte Carlo simulation, large-scale random sampling calculations were performed on the risk assessment model and the nonlinear coupling effect model to obtain the predicted probability distribution and confidence interval of the risk index. The confidence interval is compared with the preset confidence constraints. When the confidence constraints are met, the risk assessment model corresponding to the predicted probability distribution of the risk index is used as the nonlinear risk assessment model.

4. The method for safety risk classification of thin-layer weakly cemented expansive rock tunnels according to claim 3, characterized in that, The geological parameters of the expansive rock are input into a fuzzy neural network algorithm for adaptive neural fuzzy inference training to construct a risk assessment model, including: The measured data in the geological parameters of the expansive rock are used as the basic sample. The extended data in the three-dimensional numerical simulation model of geotechnical engineering and the expert experience sample data are integrated to construct a hybrid sample library. The hybrid sample library is input into the adaptive neurofuzzy inference model for network training and optimization to obtain the adaptive neurofuzzy risk assessment model. The comprehensive risk index and sensitivity analysis of the target tunnel are calculated based on the adaptive neurofuzzy risk assessment model to obtain the comprehensive risk index and sensitivity coefficient. The risk assessment model is obtained by adjusting the feedback weights of the adaptive neurofuzzy inference model based on the comprehensive risk index and the sensitivity coefficient.

5. The method for safety risk classification of thin-layer weakly cemented expansive rock tunnels according to claim 3, characterized in that, The comprehensive evaluation model of the aforementioned indicators is subjected to orthogonal numerical tests, and the surrounding rock response is assessed using a three-dimensional numerical simulation model for geotechnical engineering. A nonlinear coupling effect model is then constructed, including: The indicators affecting the stability of the surrounding rock in the comprehensive judgment model are used to design multiple sets of numerical simulation working conditions with different indicator levels through orthogonal experiments to obtain a combination of mechanical response indicators. The combination of mechanical response indices is calculated using a three-dimensional numerical simulation model for geotechnical engineering to obtain the surrounding rock stability evaluation index. The relationship between the combination of mechanical response indices and the surrounding rock stability evaluation indices is fitted and analyzed using the response surface methodology to obtain a nonlinear model. Based on regression analysis, the relationship between the surrounding rock stability evaluation index and the surrounding rock response is established, and a regression coefficient matrix is ​​constructed. By using a preset dimensionless displacement as the risk response value, and substituting the risk response value into the regression coefficient matrix to correct the nonlinear model, a nonlinear coupling effect model is obtained.

6. The method for safety risk classification of thin-layer weakly cemented expansive rock tunnels according to claim 1, characterized in that, Based on the dynamic risk threshold, tunnel construction risks are classified into levels. Combined with real-time monitoring data, a preset risk assessment model is dynamically fed back and its parameters are updated to obtain a tunnel safety risk classification scheme, including: Based on the dynamic risk threshold and model self-learning mechanism, the error between the risk prediction value and the measured deformation is calculated by real-time monitoring data to obtain the error value; When the error value exceeds a preset threshold, the incremental learning algorithm is triggered to train and update the parameters of the preset risk assessment model using real-time geological information and real-time monitoring data, thereby obtaining an updated risk assessment model. Based on the updated risk assessment model and the dynamic risk threshold, the assessment units along the tunnel are dynamically classified into risk levels to construct a tunnel safety risk classification scheme.

7. A safety risk support structure for thin-layer weakly cemented expansive rock tunnels, using the safety risk classification method for thin-layer weakly cemented expansive rock tunnels as described in any one of claims 1-6, characterized in that... include: Different risk sections in the tunnel safety risk classification scheme are configured with graded and adapted support systems. The risk sections correspond to conventional, reinforced, high-strength, and composite retractable support structures in order from lightest to highest risk.

8. The safety risk support structure for thin-layer weakly cemented expansive rock tunnels according to claim 7, characterized in that, Conventional support structures include: The first and second risk sections among the multiple risk sections are supported by ordinary advanced anchor bolts, I-beam arch frames, system mortar anchor bolts and C25 shotcrete.

9. The thin-layer weakly cemented expansive rock tunnel safety risk support structure according to claim 7, characterized in that, Reinforced support structures include: For the third-level risk section among the multiple risk sections, a combination of encrypted advanced small pipe grouting support, I-beam arch frame, extended system anchor bolts, steel mesh and locking anchor pipe, and C25 shotcrete is used for support.

10. The thin-layer weakly cemented expansive rock tunnel safety risk support structure according to claim 7, characterized in that, High-strength support structures include: For the fourth-level risk section among the multiple risk sections, pre-grouting is used for pre-support, and H-shaped steel arches, heavy I-beam steel arches, hollow grouting anchors, double-layer steel mesh and C25 shotcrete are configured. Strong locking toe structure and invert arch must be set to form a strong support structure.

Images