A soft rock TBM jamming risk intelligent prediction method based on numerical mechanism guidance

Abstract

This invention discloses an intelligent prediction method for TBM jamming risk in soft rock based on numerical mechanism guidance, comprising the following steps: S01: Establishing a refined numerical simulation model of the entire dynamic tunneling process of TBM in compressional soft rock strata. The model includes the TBM cutterhead, front, middle, and tail stage shields, soft rock, segment lining, and backfill material, simulating the evolution process; S02: Systematically calculating and analyzing surrounding rock geological parameters, TBM tunneling parameters, TBM structural parameters, and TBM construction parameters, constructing a numerical sample library of jamming indexes; S03: Predicting large deformation of soft rock and shield compression load, integrating multi-source data on TBM-rock interaction in actual engineering, and establishing an intelligent prediction model for TBM jamming risk in soft rock; S04: Assessing the TBM jamming risk and verifying the effectiveness of prevention and control measures. The technical solution of this invention can comprehensively consider the entire process of dynamic interaction between the TBM and soft rock; explain the mechanism of TBM jamming disasters; and improve the generalization ability of machine learning methods in practical applications.

Description

Technical Field

[0001] This invention relates to the field of TBM tunneling technology, and in particular to an intelligent prediction method for the risk of TBM jamming in soft rock based on numerical mechanism guidance. Background Technology

[0002] Tunnel boring machines (TBMs) are widely used in infrastructure construction projects such as water conservancy, highways, railways, and municipal transportation, especially in the construction of tunnels (caves) for long-distance water diversion and mountain railway projects. The TBM method demonstrates significant advantages over traditional tunnel construction methods in terms of construction efficiency and safety. However, as TBMs penetrate deep into rock strata, complex and variable surrounding rock conditions create a series of construction risks, including frequent geological disasters such as rock bursts, mudslides, water inrushes, large deformations due to surrounding rock compression, and the release of toxic and harmful gases. Among these, TBM jamming caused by large deformations due to soft rock compression is the most common geological disaster encountered in related projects, accounting for more than one-third. Many past and ongoing major projects have suffered from TBM jamming due to difficulty in freeing the machine, resulting in prolonged TBM shutdowns and even outright failure of the TBM method. Currently, there is no efficient method for freeing TBMs from jamming disasters; post-disaster remedial measures such as widening excavation and grouting significantly increase the workload and extend the construction period. How to predict and proactively prevent TBM jamming risks to avoid passive freeing remains a major and critical issue in the field of deep-buried TBM tunnel engineering.

[0003] Clarifying the entire process of rock-machine interaction during TBM construction and the evolution mechanism of soft rock deformation is fundamental to predicting TBM jamming risks. Currently, there are three main approaches focusing on key issues such as the deformation mechanism of weak surrounding rock and the interaction between the TBM shield and the surrounding rock. The first approach is the empirical analysis method, typically used in the engineering planning and design phases. This method predicts jamming risks based on the assessment of surrounding rock deformation. The second approach is theoretical analysis and numerical simulation calculation methods, commonly used during the construction phase to verify design parameters and analyze the causes of jamming after an accident. The third approach employs artificial intelligence methods such as machine learning for TBM jamming risk prediction.

[0004] The aforementioned methods have, to some extent, promoted the development of methods for predicting TBM jamming risks in soft rock. However, while the first two methods consider the deformation evolution mechanism of the surrounding rock in the interaction between the TBM shield and the surrounding rock, they still lack a comprehensive consideration of the pre- and post-deformation processes of the surrounding rock, such as the deformation of the cutterhead-excavation face rock mass and the lining-surrounding rock. In particular, the cutterhead excavation process (e.g., using different tunneling parameters) can have a decisive impact on the unloading stress path and damage of the weak surrounding rock, affecting the deformation evolution mechanism. The accuracy of the third method, which uses supervised machine learning methods such as Support Vector Machine (SVM), Random Forest (RF), and Decision Tree (DT), in predicting TBM jamming risks is largely limited by the size and quality of the case database and its degree of matching with actual engineering problems.

[0005] The current application status of technology in this field indicates the following shortcomings of existing technologies: First, empirical analysis, theoretical analysis, and numerical simulation methods cannot comprehensively consider the entire process of dynamic interaction between TBMs and soft rock; Second, existing data-driven machine learning methods are limited by the scale and quality of case study data, facing problems such as low matching degree of case study data and low data quality, resulting in poor machine learning performance, difficulty in generalization, and difficulty in explaining the intrinsic mechanism of TBM jamming disaster evolution. Summary of the Invention

[0006] The main objective of this invention is to address the shortcomings of the aforementioned technologies by proposing a numerical mechanism-guided intelligent prediction method for the risk of TBM jamming in soft rock formations. By integrating the advantages of the professionalism and interpretability of numerical mechanism methods with the timeliness and flexibility of data-driven intelligent prediction, this invention constructs a numerical mechanism model of the entire interaction process between TBMs and soft rock formations. This model comprehensively considers the entire dynamic interaction process between TBMs and soft rock, explains the mechanism of TBM jamming disasters, and enhances the generalization ability of machine learning methods in practical applications.

[0007] To achieve the above objectives, this invention proposes a numerical mechanism-guided intelligent prediction method for the risk of TBM (Tunnel Boring Machine) jacking in soft rock, comprising the following steps:

[0008] S01: Establish a refined numerical simulation model of the entire dynamic tunneling process of TBM in squeezing soft rock strata. The model includes TBM cutterhead, front, middle and tail stepped shields, soft rock, segment lining and backfill material behind the wall, and simulates the evolution process of soft rock rheology and damage characteristics under the action of TBM cutter rock breaking and excavation unloading.

[0009] S02: Through the numerical simulation model, the influence of surrounding rock geological parameters, TBM tunneling parameters, TBM structural parameters and TBM construction parameters on surrounding rock deformation and TBM shield compression load is systematically calculated and analyzed, and a numerical sample library of TBM shield jamming index is constructed. The jamming index includes the clearance between the surrounding rock and the shield and the intensity of the surrounding rock compression load on the shield.

[0010] S03: Based on the numerical sample library, guide the machine learning method to predict large deformation of soft rock and shield compression load, integrate multi-source data of TBM rock-machine interaction in actual engineering, and establish an intelligent prediction model for TBM jamming risk in soft rock.

[0011] S04: Use the intelligent prediction model to assess the risk of TBM card machines and verify the timing and effectiveness of the card machine proactive prevention and control measures.

[0012] Optionally, in step S01, the numerical simulation model adopts the discrete element method (DEM) to simulate the crack propagation of soft rock during the rolling cutter rock breaking process, the deformation evolution after unloading the damaged surrounding rock, and the dynamic change of the gradient gap between the front, middle and tail shields and the soft rock.

[0013] Optionally, in step S02, the surrounding rock geological parameters include the uniaxial compressive strength, strength-stress ratio, elastic modulus, and viscosity coefficient of the rock; the TBM tunneling parameters include the advance speed, cutterhead overcutting amount, and average single cutter thrust; the TBM structural parameters include the excavation diameter, shield length, and shield gradient; and the TBM construction parameters include the backfill spacing behind the tunnel lining segments and the strength of the gravel material.

[0014] Optionally, in step S03, the guided machine learning method includes:

[0015] S031: Perform a correlation test on the feature parameters in the numerical sample library and remove redundant parameters. The correlation test uses the Pearson correlation coefficient or the Spearman correlation coefficient.

[0016] S032: Train the model using logistic regression, decision tree, support vector machine, random forest or gradient boosting decision tree algorithm, and select the best prediction model through cross-validation;

[0017] S033: The fused TBM rock-machine interaction multi-source data includes TBM tunneling characteristic parameters, shield friction resistance, and measured data from shield sensors.

[0018] Optionally, in step S033, the shield friction resistance is obtained by back-calculation through regression analysis of the total thrust and penetration characteristic values ​​of the empty thrust section, loading section and stabilization section of the complete TBM tunneling cycle.

[0019] Optionally, in step S03, the intelligent prediction model performs card risk assessment in the following manner:

[0020] S61: When the predicted surrounding rock deformation exceeds the gap outside the shield, the first early warning indicator is triggered;

[0021] S62: Based on the first early warning indicator, predict the correlation between the shield's compressive load and the long-term deformation characteristics of the surrounding rock, and obtain the second early warning indicator;

[0022] S63: Combine the maximum thrust reserve of the TBM to comprehensively determine the risk level of engine jamming.

[0023] Optionally, in step S04, the card machine proactive prevention measures include:

[0024] S041: Optimize tunneling parameters, including adjusting advance speed, penetration depth, and radial overcut;

[0025] S042: Take measures such as ground pretreatment, adjust the timing of backfilling behind the tunnel lining segments, or strengthen the strength of the tunnel lining segments and backfill materials.

[0026] Optionally, in step S01, the numerical simulation model also considers the creep / softening characteristics of soft rock, and simulates the age-related softening phenomenon of rock particles under maximum tensile stress by constructing a micro-bonded contact model.

[0027] The beneficial effects of the technical solution of this invention are:

[0028] 1. By establishing a refined numerical model of the entire dynamic tunneling process of TBM in squeezing soft rock strata, the system calculates the influence of multiple parameters on the deformation of the surrounding rock and the squeezing load of the shield and constructs a numerical sample library. Then, it guides machine learning to integrate multi-source data from actual engineering projects to establish an intelligent prediction model for TBM jamming risk. Finally, it realizes risk assessment and verification of prevention and control measures, and achieves dynamic simulation of the entire process of soft rock rheology and damage evolution under TBM excavation unloading. This solves the problem that existing technologies cannot comprehensively consider the entire process of rock-machine dynamic interaction.

[0029] 2. This method integrates mechanistic knowledge with multi-source data, generates a high-quality sample library across working conditions through numerical simulation, improves the generalization ability of machine learning methods in practical applications, explains the mechanism of TBM truck malfunction, and overcomes the model training problem in existing technologies that lack big data on actual engineering TBM truck malfunctions. Attached Figure Description

[0030] Figure 1 This is a flowchart illustrating an embodiment of the prediction method of the present invention;

[0031] Figure 2 This is a schematic diagram of a refined discrete element model for numerical simulation of the entire soft rock TBM construction process in an embodiment of the prediction method of the present invention.

[0032] Figure 3 This is a schematic diagram of the soft rock TBM cutter rock breaking process model calculation in an embodiment of the prediction method of the present invention;

[0033] Figure 4 A statistical graph showing the results of Spearman correlation coefficient test on the feature values ​​of numerical samples in an embodiment of the prediction method of the present invention.

[0034] Figure 5 This is a comparison chart of the machine learning prediction values ​​and numerical mechanism calculation values ​​of the TBM card machine index in this prediction method embodiment.

[0035] The following are the symbols and their meanings: 1. TBM cutterhead; 2. Front shield; 3. Middle shield; 4. Tail shield; 5. Segment; 6. Gravel filling layer; 7. Soft rock; 8. Model boundary constraint; 11. First disc-shaped cutter ring; 12. Second disc-shaped cutter ring; 71. Rock fragment; 72. Rock mass fissure. Detailed Implementation

[0036] 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 a part of the embodiments of the present invention, and not all of them. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0037] This invention proposes an intelligent prediction method for the risk of TBM trucks in soft rock, guided by numerical mechanisms.

[0038] In embodiments of the present invention, such as Figures 1 to 5 As shown, the intelligent prediction method for TBM truck risk in soft rock based on numerical mechanism includes the following steps:

[0039] S01: Establish a refined numerical simulation model of the entire dynamic tunneling process of TBM in squeezing soft rock strata. The model includes TBM cutterhead, front, middle and tail stepped shields, soft rock, segment lining and backfill material behind the wall, and simulates the evolution process of soft rock rheology and damage characteristics under the action of TBM cutter rock breaking and excavation unloading.

[0040] S02: Through numerical simulation models, systematically calculate and analyze the influence of surrounding rock geological parameters, TBM tunneling parameters, TBM structural parameters and TBM construction parameters on surrounding rock deformation and TBM shield compression load, and construct a numerical sample library of TBM shield jamming indexes. The jamming indexes include the clearance margin between the surrounding rock and the shield and the intensity of the surrounding rock compression load on the shield.

[0041] S03: Based on numerical sample library-guided machine learning methods, predict large deformation of soft rock and shield compression load, integrate multi-source data of TBM rock-machine interaction in actual engineering, and establish intelligent prediction model for TBM jamming risk in soft rock.

[0042] S04: Use intelligent predictive models to assess the risk of TBM malfunctions and verify the timing and effectiveness of proactive malfunction prevention measures.

[0043] This invention establishes a refined numerical model of the entire dynamic tunneling process of a TBM in squeezing soft rock strata. It systematically calculates the influence of multiple parameters on surrounding rock deformation and shield squeezing load, constructs a numerical sample library, and then guides machine learning to integrate multi-source data from actual engineering projects to establish an intelligent prediction model for TBM jamming risks. Ultimately, it achieves risk assessment and verification of prevention and control measures, realizing a dynamic simulation of the entire process of soft rock rheology and damage evolution under TBM excavation and unloading. It quantitatively reveals the interaction mechanism of geological, tunneling, and structural parameters, solving the problem that existing technologies cannot comprehensively consider the dynamic interaction between rock and TBM. Through numerical simulation, a high-quality sample library is generated across working conditions, combined with… Correlation testing eliminates redundant features, overcoming the bottleneck of scarce actual machine jam data leading to difficulties in machine learning training and reducing model generalization error. The intelligent prediction model, which integrates numerical mechanisms and real-time monitoring data, adopts a dual-indicator early warning system of "deformation + load," significantly improving the accuracy of machine jam warnings compared to traditional single-criteria systems, and the results can be traced back to the specific parameter influence mechanism. The timing of prevention and control measures and parameter optimization, such as pre-grouting and tunneling parameter adjustment, are verified through numerical models, forming a "prediction-intervention-verification" closed loop. This reduces the risk level of machine jams while improving construction efficiency, significantly enhancing the safety and economy of soft rock TBM construction.

[0044] In some embodiments, the numerical simulation model employs the discrete element method (DEM) to simulate the crack propagation of soft rock during the rolling cutter rock breaking process, the deformation evolution after unloading of damaged surrounding rock, and the dynamic changes in the gradient gap between the front, middle, and tail shields and the soft rock.

[0045] Specifically, the numerical simulation model constructed using the Discrete Element Method (DEM) can subdivide soft rock into discrete units, accurately simulating the entire process of crack initiation, propagation, and convergence during rock breaking by the cutter head at the microscopic level, and intuitively presenting the rock fragmentation mechanism. Simultaneously, for the deformation evolution of the damaged surrounding rock after unloading, the deformation amount and rate of the surrounding rock can be quantified by tracking the displacement and contact force changes between units. Furthermore, for the dynamic changes in the gradient gap between the front, middle, and tail shields and the soft rock, the DEM can capture the changes in the contact state between the surrounding rock and the shield during shield advancement in real time, accurately calculating the gap margin. This lays the foundation for the accurate acquisition of TBM shield jamming indicators (such as gap margin and compression load), comprehensively improving the fidelity and prediction accuracy of numerical simulation of the complex mechanical behavior of TBM tunneling in soft rock.

[0046] In some embodiments, the surrounding rock geological parameters include the uniaxial compressive strength, strength-stress ratio, elastic modulus, and viscosity coefficient of the rock; the TBM tunneling parameters include the advance speed, cutterhead overcutting, and average single-cutter thrust; the TBM structural parameters include the excavation diameter, shield length, and shield gradient; and the TBM construction parameters include the backfill spacing behind the tunnel lining segments and the strength of the gravel material.

[0047] Specifically, by systematically incorporating geological parameters such as uniaxial compressive strength and strength-stress ratio of rock, the fundamental impact of soft rock mechanical properties on surrounding rock deformation can be quantified. Introducing tunneling parameters such as advance speed and cutterhead over-excavation allows analysis of the time-dependent effects of stress release rate and surrounding rock exposure time on deformation during dynamic construction. Combining structural parameters such as excavation diameter and shield gradient reveals the regulatory mechanism of shield geometry on the distribution of contact pressure in the surrounding rock. Incorporating construction parameters such as backfill spacing behind the tunnel lining segments allows for the assessment of the constraint effect of support measures on surrounding rock deformation. The synergistic analysis of various parameters not only comprehensively covers the multi-dimensional influencing factors of "geology-equipment-construction" and quantifies the interaction mechanism between parameters, but also provides high-quality training data for machine learning across scenarios by constructing a numerical sample library containing multiple working conditions and parameter combinations. This effectively solves the problem of insufficient model generalization ability caused by the scarcity of machine-related data in actual engineering, enabling subsequent intelligent prediction models to more accurately capture the evolution of surrounding rock deformation and shield compression load during soft rock TBM tunneling.

[0048] In some embodiments, the guided machine learning method includes:

[0049] S031: Perform correlation tests on the feature parameters in the numerical sample library, and remove redundant parameters. The correlation test uses Pearson correlation coefficient or Spearman correlation coefficient.

[0050] S032: Train the model using logistic regression, decision tree, support vector machine, random forest or gradient boosting decision tree algorithm, and select the best prediction model through cross-validation;

[0051] S033: The fused TBM rock-machine interaction multi-source data includes TBM tunneling characteristic parameters, shield friction resistance, and measured data from shield sensors.

[0052] Specifically, by eliminating redundant parameters in the numerical sample database through correlation tests (Pearson / Spearman coefficients), the feature dimensionality can be significantly reduced, the computational complexity of model training can be decreased, and the interference of the "curse of dimensionality" on prediction accuracy can be avoided. At the same time, core parameters that significantly affect the risk of jamming (such as strength-stress ratio, shield gradient, etc.) are retained, making the model more focused on key influencing factors. By using multiple machine learning algorithms (logistic regression, random forest, etc.) combined with cross-validation to select the best model, the advantages of different algorithms can be utilized (such as the interpretability of decision trees and the nonlinear fitting ability of gradient boosting trees), while avoiding the limitations of a single algorithm through multiple rounds of validation. Limitations are addressed by selecting the prediction model with the smallest generalization error, thus improving adaptability to complex working conditions. By integrating multi-source information such as TBM tunneling characteristic parameters, shield friction, and sensor measured data, the mechanism analysis of numerical simulation can be deeply coupled with real-time on-site monitoring data. This compensates for the limitations of single numerical simulation in terms of working conditions and the scarcity of actual data samples, forming a dual constraint of "mechanism guidance + data drive". This enables the model to capture the physical essence of soft rock rheology and adapt to the dynamic changes on the construction site, ultimately achieving accurate early warning of "deformation + load" dual indicators. This significantly improves the timeliness and reliability of machine jam risk prediction compared to traditional single criteria.

[0053] In some embodiments, the shield friction is calculated by back-calculating the total thrust and penetration characteristic values ​​of the TBM tunneling cycle in the empty thrust section, loading section and stabilization section.

[0054] Specifically, by performing regression analysis on the total thrust and penetration characteristics of the TBM tunneling cycle in the empty-push, loading, and stabilization stages to calculate the shield friction resistance, this method can accurately isolate interfering factors such as rock-breaking thrust and equipment weight, and extract the friction load between the surrounding rock and the shield. Based on the mechanical characteristics of each stage of tunneling (the empty-push stage contains only shield friction, the loading stage is superimposed with rock-breaking load, and the stabilization stage forms a dynamic equilibrium), this method eliminates irrelevant variables through regression modeling, making the calculated friction resistance closer to the actual load state on site. Integrating real-time friction resistance data into a machine learning model can compensate for the assumptions of numerical simulation regarding complex geology, forming a "load-deformation" dual-dimensional constraint with deformation monitoring data, accurately capturing the dynamic evolution of rock-machine interaction. As a direct mechanical indicator of the shield's compression state, the real-time calculation results of friction resistance, together with the clearance margin, can form a "deformation-load" dual-indicator early warning system, significantly improving the timeliness and interpretability of the jamming risk prediction mechanism, and providing a more reliable mechanical basis for the parameter optimization of advanced prevention and control measures.

[0055] In some embodiments, the intelligent prediction model achieves SIM card risk assessment in the following ways:

[0056] S61: When the predicted surrounding rock deformation exceeds the gap outside the shield, the first early warning indicator is triggered;

[0057] S62: Based on the first early warning indicator, predict the correlation between the shield's compressive load and the long-term deformation characteristics of the surrounding rock, and obtain the second early warning indicator;

[0058] S63: Combine the maximum thrust reserve of the TBM to comprehensively determine the risk level of engine jamming.

[0059] Specifically, S61 uses the deformation of the surrounding rock exceeding the gap outside the shield as the first early warning indicator, directly quantifying the physical critical state of the surrounding rock squeezing the shield, and ensuring the rapid identification of immediate risks.

[0060] Based on the first early warning, S62 further analyzes the correlation between the compressive load and the long-term rheological characteristics of the surrounding rock, captures the continuous squeezing effect caused by soft rock creep, and avoids misjudgment of long-term risks by a single deformation index.

[0061] S63 incorporates the maximum thrust reserve parameters of TBM into the risk assessment system, and can accurately predict the risk level of jamming when the extrusion load approaches the thrust limit of the equipment.

[0062] This assessment mechanism, through a multi-level logic of "deformation critical value early warning + load-rheology correlation analysis + equipment dynamic margin verification," forms a three-dimensional assessment system covering "instantaneous state - evolution trend - equipment adaptability." Compared with traditional single deformation or load criteria, it can not only capture the progressive machine jamming risk caused by soft rock rheology in advance, but also provide an operable risk level judgment based on equipment performance parameters. This provides a quantitative basis for the precise implementation of prevention and control measures such as advanced grouting and tunneling parameter adjustment, significantly improving the timeliness and engineering guidance of risk prediction.

[0063] In some embodiments, the measures to prevent and control the jamming of the tunneling machine include: S041: optimizing tunneling parameters, including adjusting the advance speed, penetration depth and radial over-excavation; S042: taking measures such as stratum pretreatment, adjusting the timing of backfilling behind the tunnel lining segments, or strengthening the strength of the tunnel lining segments and backfill materials.

[0064] Specifically, in S041, adjusting the advance speed can dynamically match the rheological rate of soft rock, avoiding stress concentration caused by excessively rapid advance; controlling the penetration depth can reduce the range of rock disturbance caused by a single rock breaking; increasing the radial over-excavation amount buffers the deformation of the surrounding rock by reserving additional gaps. These three measures work together to reduce the risk of direct compression between the shield and the surrounding rock. In S042, pretreatment such as advanced grouting can improve the integrity of the surrounding rock and reduce the amount of rheological deformation; accurately controlling the backfilling timing behind the tunnel lining segment (such as shortening the backfilling step distance) can promptly generate support reaction force and inhibit the continuous convergence of the surrounding rock; strengthening the strength of the tunnel lining segment and backfill material can improve the compression resistance of the support structure and constrain the deformation of the surrounding rock within a safe threshold. This progressive intervention, which involves "dynamic adaptation of construction parameters, improvement of geological mechanical properties, and enhancement of support structure stiffness," not only allows for real-time adjustment of tunneling strategies based on early warning results from intelligent prediction models, but also fundamentally improves the rock-machine interaction conditions. Compared to traditional post-event emergency treatment, it can reduce the risk of machine jamming by more than 60%, while simultaneously increasing TBM construction efficiency by 15%-20% by reducing downtime for adjustment, thus achieving a dual optimization of safety and efficiency in soft rock tunnel excavation.

[0065] In some embodiments, the numerical simulation model also considers the creep / softening characteristics of soft rock, and simulates the age-related softening of rock particles under maximum tensile stress by constructing a micro-bonded contact model.

[0066] Specifically, by considering the creep / softening characteristics of soft rock in the numerical simulation model and constructing a micro-bonded contact model to simulate the time-dependent softening phenomenon of rock particles under maximum tensile stress, the strength decay and deformation accumulation characteristics of soft rock over time can be accurately characterized from the perspective of micromechanical mechanisms. On the one hand, it can capture in real time the continuous deformation of soft rock caused by creep and the dynamic growth process of shield compression load after TBM excavation and unloading, which is closer to the coupled evolution law of "deformation-time-stress" of soft rock in the field than the traditional elastic model. On the other hand, by simulating the degradation process of interparticle bond strength over time, the weakening effect of time-dependent softening of soft rock on the self-stabilizing ability of surrounding rock can be quantitatively revealed, thereby more accurately calculating the gradient gap change and compression load distribution between the shield and the surrounding rock. The introduction of this mechanism enables numerical models to output "deformation-load" evolution data that includes the time dimension, providing mechanistic support for the subsequent construction of jamming risk indicators that consider time effects (such as long-term deformation and time-varying extrusion load). This allows intelligent prediction models to more accurately predict the progressive jamming risk caused by soft rock creep, providing quantitative basis for the timing selection of advanced prevention and control measures (such as grouting before the creep acceleration stage), and significantly improving the timeliness of risk prediction and the pertinence of prevention and control measures.

[0067] Example 1

[0068] Step S01, as follows Figure 2 , Figure 3As shown, a refined numerical simulation model of the entire dynamic tunneling process of TBM in squeezing soft rock strata is established.

[0069] The numerical simulation model for TBM tunneling in soft rock formations established in this invention includes a TBM cutterhead 1, a front shield 2, a middle shield 3, a tail shield 4, tunnel segments 5, a gravel filling layer 6, soft rock 7, model boundary constraints 8, a first disc-shaped cutter ring 11, a second disc-shaped cutter ring 12, rock fragments 71, and rock fissures 72. This model can simulate in detail the process of rock fissures 72 and rock fragments 71 forming inside the soft rock 7 under the loading action of the first disc-shaped cutter ring 11 and the second disc-shaped cutter ring 12. The front shield 2, middle shield 3, and tail shield 4 have different widths of spacing with the soft rock 7 outside them to simulate the gradient distribution of the TBM shield. After the cutterhead 1 excavates the soft rock 7, it continues to tunnel forward. The surrounding rock outside the TBM shield is damaged and deformed. When the deformation exceeds the excavation gap, the front shield 2 and / or the middle shield 3 and / or the tail shield 4 come into contact with the soft rock 7 and begin to transmit the compressive load. In addition, this model also considers the support effect of segment 5 and gravel filling layer 6. This model can simulate the rock-machine interaction throughout the entire TBM tunneling process and obtain quantitative indicators such as soft rock deformation curves and shield compression load strength.

[0070] Step S02: Construct a numerical sample library of TBM shield tunneling indicators. In this embodiment, the discrete element method (DEM) is selected to simulate the entire process of dynamic tunneling of TBM in soft rock with large compression deformation characteristics.

[0071] High-stress soft rock is a common geological condition encountered during deep-buried tunnel TBM construction, easily inducing large deformations due to surrounding rock compression. Under the stress unloading and disturbance damage during TBM excavation, the weak surrounding rock undergoes significant plastic deformation. When the deformation exceeds the gap outside the shield, the first warning indicator is triggered, and the surrounding rock begins to compress the TBM shield. This study predicts the correlation between the shield's compressive load level and the characteristics of the surrounding rock, obtaining the second key indicator for TBM jamming. Based on soft rock mechanical property tests, this embodiment constructs a creep / softening bond contact model with soft rock rheological characteristics to simulate the material properties of soft rock. This microscopic contact model can simulate the softening / creep phenomenon that occurs over time after the rock particles reach maximum tensile stress.

[0072] The calculation parameters considered in this embodiment are as follows: the geological parameters of the surrounding rock include the uniaxial compressive strength, strength-stress ratio, elastic modulus and viscosity coefficient of the rock; the TBM tunneling parameters include the advance speed, cutterhead over-excavation amount and average single cutter thrust; the TBM structural parameters include the excavation diameter, shield length and shield gradient; the TBM construction parameters include the backfill step distance behind the segment wall and the strength of the gravel material.

[0073] The deformation evolution law of surrounding rock under different levels of influencing factors was calculated, and numerical samples driving machine learning were obtained.

[0074] Correlation test of numerical sample eigenvalues. Since there are numerous parameters influencing surrounding rock deformation during TBM tunneling, Pearson correlation coefficient and Spearman's rank correlation coefficient will be used to evaluate the correlation between various eigenvalues. This ensures that the selected eigenvalues ​​meet the independence requirements and minimizes the spatial dimensionality of the prediction model. Figure 3 As shown.

[0075] Step S03: Based on the numerical sample library, guide the machine learning method to predict the large deformation of soft rock and the shield compression load, and integrate multi-source data of TBM rock-machine interaction in actual engineering to establish an intelligent prediction model for the risk of TBM jamming in soft rock.

[0076] like Figure 4 As shown, the correlation test of numerical sample feature values ​​uses Pearson correlation coefficient and Spearman's rank correlation coefficient to evaluate the correlation between each feature parameter, ensuring that the feature parameter selection meets the independence requirement and minimizing the spatial dimension of the prediction model.

[0077] Model training, validation, and feature importance analysis involve comparing the prediction accuracy of numerical sample training and test sets using methods such as logistic regression (LR), decision tree (DT), support vector machine (SVM), random forest (RF), and gradient boosting decision tree (XGBoost / LightGBM). After selecting the optimal prediction model, cross-validation is performed, and the importance of each feature parameter is analyzed.

[0078] like Figure 5 As shown, model validation was performed. In a real-world single-shield TBM construction section, suitable tunneling parameters were selected for a section with a risk of machine jamming. Machine learning methods were used to calculate the surrounding rock deformation and the compressive load of the surrounding rock on the shield. Sensor data pre-deployed on the shield was extracted to validate the accuracy of the prediction results.

[0079] TBM rock-mechanical interaction multi-source data acquisition includes the following steps:

[0080] ① Extraction of TBM tunneling characteristic parameters. TBM tunneling parameters in multiple sections of squeezing surrounding rock were selected, and feature values ​​were extracted as input parameters for the prediction model.

[0081] ② Extraction of TBM tunneling friction resistance. The friction resistance of a TBM varies at different locations during tunneling. Unlike existing methods that assume a constant friction coefficient to calculate the friction resistance of the shield, this embodiment uses regression analysis of tunneling parameters to accurately obtain the friction resistance of the current TBM tunneling cycle. The key technical point is to extract the characteristic parameters of the empty thrust phase, loading phase, and stabilization phase experienced by the TBM in a complete tunneling cycle, perform regression analysis on the characteristic values ​​of total thrust and penetration depth, and obtain the total friction resistance currently experienced by the TBM based on the fitting relationship. Furthermore, the friction coefficient between the shield and the surrounding rock is calculated from the total friction resistance.

[0082] Step S04: Use the intelligent prediction model to assess the risk of TBM malfunction and verify the timing and effectiveness of the proactive prevention and control measures for malfunctions.

[0083] Intelligent risk assessment of TBM jamming. This involves integrating multi-source data on TBM tunneling parameters, geological parameters, and support parameters to predict surrounding rock deformation and shield load, and analyzing the jamming risk in conjunction with the TBM's maximum thrust reserve.

[0084] Based on the risk assessment of TBM (Toyota Machine) jamming, the following prevention and control measures for TBM construction are proposed, including the following aspects:

[0085] ① Tunneling parameter optimization based on TBM jamming risk assessment. Based on the current tunneling parameters and geological parameters of the TBM in actual engineering examples, machine learning is used to predict the deformation of the surrounding rock and the shield compression load level. When a potential jamming risk section is predicted, the current tunneling parameter status is used to calculate whether the TBM shield can pass smoothly. The main control tunneling parameters such as advance speed, penetration depth, and radial over-excavation are optimized, and the shutdown timing is optimized according to the evolution law of surrounding rock deformation and the geometric characteristics of the TBM shield.

[0086] ② Proactive prevention and control of TBM jamming. The model established in step S01 is used to calculate the control effect of proactive prevention and control measures on reducing the risk of TBM jamming, including the characteristics and timing of pre-grouting reinforcement of the strata, TBM segment reinforcement, gravel, and backfill grouting materials. The characteristic values ​​of the prevention and control scheme for the disaster-causing factors of TBM jamming are input into the TBM jamming risk prediction model to verify the effectiveness of the prevention and control measures.

[0087] The above description is merely a preferred embodiment of the present invention and does not limit the patent scope of the present invention. Any equivalent structural transformations made using the contents of the present invention's specification and drawings under the inventive concept of the present invention, or direct / indirect applications in other related technical fields, are included within the patent protection scope of the present invention.

Claims

1. A numerical mechanism-guided intelligent prediction method for the risk of TBM trucks in soft rock, characterized in that, Includes the following steps: S01: Establish a refined numerical simulation model of the entire dynamic tunneling process of TBM in squeezing soft rock strata. The model includes the TBM cutterhead, front, middle and tail stepped shields, soft rock, segment lining and backfill material behind the wall, and simulates the evolution of the rheological and damage characteristics of soft rock under the action of TBM cutter rock breaking and excavation unloading. The model simulates the crack propagation of soft rock during the cutter rock breaking process, the deformation evolution after unloading of damaged surrounding rock, and the dynamic change of the gradient gap between the front, middle and tail stepped shields and soft rock. S02: Through the numerical simulation model, the influence of surrounding rock geological parameters, TBM tunneling parameters, TBM structural parameters and TBM construction parameters on surrounding rock deformation and TBM shield compression load is systematically calculated and analyzed, and a numerical sample library of TBM shield jamming index is constructed. The jamming index includes the clearance between the surrounding rock and the shield and the intensity of the surrounding rock compression load on the shield. S03: Based on the numerical sample library, a machine learning method is used to predict large deformations in soft rock and shield compression loads. Multi-source data on TBM rock-machine interaction in actual engineering is integrated to establish an intelligent prediction model for TBM jamming risks in soft rock. The integrated TBM rock-machine interaction multi-source data includes TBM tunneling characteristic parameters, shield friction resistance, and measured data from shield sensors. The shield friction resistance is calculated by back-calculating the total thrust and penetration characteristic values ​​of the empty thrust section, loading section, and stabilization section of the complete TBM tunneling cycle. S04: Use the intelligent prediction model to assess the risk of TBM card machines and verify the timing and effectiveness of the card machine proactive prevention and control measures.

2. The method as described in claim 1, characterized in that, In step S01, the numerical simulation model adopts the discrete element method.

3. The method as described in claim 1, characterized in that, In step S02, the surrounding rock geological parameters include the uniaxial compressive strength, strength-stress ratio, elastic modulus, and viscosity coefficient of the rock; the TBM tunneling parameters include the advance speed, cutterhead over-excavation amount, and average single-cutter thrust; the TBM structural parameters include the excavation diameter, shield length, and shield gradient; and the TBM construction parameters include the backfill spacing behind the segment wall and the strength of the gravel material.

4. The method as described in claim 1, characterized in that, In step S03, the guided machine learning method includes: S031: Perform a correlation test on the feature parameters in the numerical sample library and remove redundant parameters. The correlation test uses the Pearson correlation coefficient or the Spearman correlation coefficient. S032: Use logistic regression, decision tree, support vector machine or random forest, and select the best prediction model through cross-validation; S033: The fused TBM rock-machine interaction multi-source data includes TBM tunneling characteristic parameters, shield friction resistance, and measured data from shield sensors.

5. The method as described in claim 1, characterized in that, In step S03, the intelligent prediction model achieves card machine risk assessment in the following way: S61: When the predicted surrounding rock deformation exceeds the gap outside the shield, the first early warning indicator is triggered; S62: Based on the first early warning indicator, predict the correlation between the shield's compressive load and the long-term deformation characteristics of the surrounding rock, and obtain the second early warning indicator; S63: Combine the maximum thrust reserve of the TBM to comprehensively determine the risk level of engine jamming.

6. The method as described in claim 1, characterized in that, In step S04, the advanced prevention and control measures for the card machine include: S041: Optimize tunneling parameters, including adjusting advance speed, penetration depth, and radial overcut; S042: Take measures such as ground pretreatment, adjust the timing of backfilling behind the tunnel lining segments, or strengthen the strength of the tunnel lining segments and backfill materials.

7. The method as described in claim 1, characterized in that, In step S01, the numerical simulation model also considers the creep / softening characteristics of soft rock, and simulates the age-related softening phenomenon of rock particles under maximum tensile stress by constructing a micro-bonded contact model.

Images