Tunneling machine risk early warning method, system, device and storage medium

Abstract

This application provides a method, system, device, and storage medium for early warning of tunnel boring machine (TBM) jamming risks, relating to the field of tunnel boring technology. The method includes: determining multiple jamming event types and corresponding preset jamming risk levels based on historical event data; identifying risk indicators and quantitative standards for each jamming event type through geological risk source identification; acquiring multi-source heterogeneous data from the TBM and performing data preprocessing to obtain cutterhead feature data and target feature data; calculating the error rate based on the cutterhead feature data and target cutterhead threshold, and determining the risk probability level of each jamming event type based on the error rate; calculating the dynamic risk assessment weight of each risk indicator in each jamming event type, thereby determining the jamming risk level; and jointly determining the risk warning level for each jamming event type based on the risk probability level and the jamming risk level. This application improves the robustness and decision-making credibility of early warning for TBM jamming risks.

Description

Technical Field

[0001] This application relates to the field of tunnel boring technology, specifically to a method, system, device, and storage medium for early warning of machine jamming risk in tunnel boring machines. Background Technology

[0002] When carrying out tunnel excavation in the complex geological environment of plateau, the prediction of tunnel boring machine (TBM) jamming risk faces significant challenges. Current mainstream methods mainly rely on qualitative assessments from geological survey reports or use single parameter thresholds such as thrust and torque exceeding limits for early warning.

[0003] However, these linear process-based methods struggle to systematically consider the dynamic interactive coupling effects of multiple disaster factors such as high ground stress, large deformation of soft rock, and rockbursts, and lack the ability to effectively fuse and analyze multi-source heterogeneous data. These limitations directly lead to insufficient early warning timeliness and poor assessment accuracy, failing to meet the practical engineering needs for accurate quantitative assessment of jamming risks under complex working conditions. Summary of the Invention

[0004] In view of the above-mentioned shortcomings of the prior art, this application provides a method, system, equipment and storage medium for early warning of tunnel boring machine jamming risk, which effectively solves the problem of insufficient accuracy in quantitative assessment of tunnel boring machine jamming risk under complex geological conditions in plateau areas.

[0005] Firstly, this application provides a method for early warning of jamming risk in a tunnel boring machine, the method comprising: Based on historical event data, tunnel boring machine jamming event types are classified to determine multiple jamming event types and corresponding preset jamming risk levels; Geological risk sources are identified for the target construction section of the tunnel boring machine, and risk indicators and quantitative standards for each type of machine jamming event are determined. Acquire multi-source heterogeneous data of the tunnel boring machine, perform data preprocessing on the multi-source heterogeneous data, and obtain cutterhead feature data and target feature data corresponding to the risk indicators; The error rate is calculated based on the cutter head feature data and the target cutter head threshold, and the risk probability level of each of the jamming event types is determined based on the error rate. Calculate the dynamic risk assessment weight of each risk indicator in each of the card-operated event types, and determine the card-operated risk level of each of the card-operated event types based on the dynamic risk assessment weight and the target feature data; The risk warning level for each type of card-related event is determined by combining the risk probability level and the card-related risk level.

[0006] In an optional implementation, calculating the dynamic risk assessment weight of each risk indicator in each of the card event types includes: Determine the best and worst indicators for each of the aforementioned risk indicators, and calculate the target subjective weight for each of the aforementioned risk indicators based on the best and worst indicators; Calculate the index entropy value based on the risk index data of each risk index in historical card events, and determine the target objective weight of each risk index based on the index entropy value; The target subjective weight and target objective weight of each risk indicator are weighted and combined to obtain the combined weight of each risk indicator; The combined weights are dynamically adjusted based on the contribution scores of each risk indicator to the card machine risk, thereby obtaining the dynamic risk assessment weights of each risk indicator.

[0007] In an optional implementation, determining the card-operated risk level for each of the card-operated event types based on the dynamic risk assessment weights and the target feature data includes: Based on the risk indicators and the preset card machine risk level, construct a minimum value standard matrix and a maximum value standard matrix within a preset value range; Calculate multiple single-index attribute measurement matrices based on the target feature data, the minimum value standard matrix, and the maximum value standard matrix; The individual single-indicator attribute measurement vectors are modified based on the risk coupling coefficients between the individual risk indicators to obtain the modified individual single-indicator attribute measurement vectors. Each of the modified single-indicator attribute measurement matrices is fused with the dynamic risk assessment weight to obtain the corresponding multi-indicator attribute measurement vector. The target attribute measure vector is obtained by calculating the mean of each of the multi-index attribute measure vectors. The risk quantification value is calculated based on the target attribute measurement vector and the maximum value of the endpoints of each measurement interval, and the card-freezing risk level of each card-freezing event type is determined based on the risk quantification value.

[0008] In an optional implementation, determining the best and worst indicators for each of the risk indicators, and calculating the target subjective weight of each of the risk indicators based on the best and worst indicators, includes: Determine the set of evaluation indicators for each of the aforementioned risk indicators, and determine the best and worst indicators based on the set of evaluation indicators. Calculate the importance ratio of the optimal indicator and each evaluation indicator in the set of evaluation indicators to obtain the first comparison vector; Calculate the importance ratio of the worst-case indicator and each of the evaluation indicators in the set of evaluation indicators to obtain a second comparison vector; A subjective weight optimization function is constructed based on the best indicator, the worst indicator, the first comparison vector, and the second comparison vector. Set the boundary conditions for the subjective weight optimization function, calculate the subjective weight when the subjective weight optimization function reaches its minimum value, and obtain the target subjective weight.

[0009] In an optional implementation, the entropy value of each risk indicator is calculated based on the risk indicator data in historical card-operated events, and the target objective weight of each risk indicator is determined based on the entropy value, including: Obtain risk indicator data for each of the aforementioned risk indicators under multiple historical card-operated events, and construct an indicator data matrix based on each of the aforementioned risk indicators and the indicator data; The index data matrix is ​​standardized to obtain a standardized data matrix; Normalize each element of the standardized data matrix to obtain a standard normalized data matrix; The index entropy value of each risk indicator is calculated based on the standard normalized data matrix, and the objective weight is calculated based on the index entropy value to obtain the target objective weight.

[0010] In an optional implementation, the step of dynamically adjusting the combined weights based on the contribution scores of each risk indicator to the card machine risk, to obtain the dynamic risk assessment weights of each risk indicator, includes: Construct an indicator state vector based on the contribution score of each of the aforementioned risk indicators, and determine the balance function of the indicator state vector; The dynamic risk assessment weights are obtained by calculating the combined weights of each risk indicator and the normalized product of the indicator state vectors.

[0011] In an optional implementation, the step of calculating the error rate based on the cutter head feature data and cutter head threshold data, and determining the risk probability level of each of the jamming event types based on the error rate, includes: A tool turret threshold prediction model is constructed, and the tool turret threshold data is predicted based on the tool turret feature data and the target feature data using the tool turret threshold prediction model. The error rate is calculated based on the cutter head feature data and the cutter head threshold data, and the risk probability level is determined based on the error rate and the preset error threshold.

[0012] Secondly, this application provides a jamming risk early warning system for a tunnel boring machine, the system comprising: The machine jamming event classification module is used to classify the types of machine jamming events of tunnel boring machines based on historical event data, and to determine multiple machine jamming event types and corresponding preset machine jamming risk levels; The risk indicator determination module is used to identify geological risk sources in the target construction section of the tunnel boring machine and determine the risk indicators and quantitative standards for each type of machine jamming event. The data acquisition and processing module is used to acquire multi-source heterogeneous data of the tunnel boring machine, perform data preprocessing on the multi-source heterogeneous data, and obtain cutterhead feature data and target feature data corresponding to the risk indicators. The first risk prediction module is used to calculate the error rate based on the cutter head feature data and the target cutter head threshold, and to determine the risk probability level of each of the jamming event types based on the error rate. The second risk prediction module is used to calculate the dynamic risk assessment weight of each risk indicator in each of the card machine event types, and determine the card machine risk level of each of the card machine event types based on the dynamic risk assessment weight and the target feature data. The risk warning determination module is used to jointly determine the risk warning level of each of the card machine event types based on the risk probability level and the card machine risk level.

[0013] Thirdly, this application provides an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the tunnel boring machine jam risk warning method as described in the first aspect of this application.

[0014] Fourthly, this application provides a computer-readable storage medium having a computer program stored thereon, wherein the computer program, when executed by a processor, implements the tunnel boring machine jam risk warning method as described in the first aspect of this application.

[0015] This application provides a method, system, equipment, and storage medium for early warning of tunnel boring machine (TBM) jamming risks. It constructs a refined classification system for jamming event types, enabling early warnings to accurately match risk scenarios with different geological causes, significantly improving the targeting and engineering adaptability of jamming risk warnings. Simultaneously, it introduces multi-risk factor coupling mechanism analysis, breaking through the linear limitations of traditional single-parameter thresholds and enhancing the predictive ability for sudden and complex jamming events. Relying on an intelligent sensor network and hierarchical perception architecture, it achieves closed-loop data support from long-distance macroscopic prediction to short-distance dynamic response, ensuring the timeliness and on-site verifiability of risk assessment. Combining a dynamic weighting mechanism and an improved attribute interval model, the risk warning indicators can adaptively adjust with the real-time evolution of the surrounding rock condition, comprehensively improving the robustness and decision-making reliability of the TBM early warning system under complex and variable geological conditions in plateau regions. Attached Figure Description

[0016] To more clearly illustrate the technical solutions of the embodiments of this application, 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 this application and should not be regarded as a limitation of 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 tunnel boring machine jam risk warning method provided in the embodiments of this application; Figure 2 This is a schematic diagram of the jamming risk warning system for a tunnel boring machine provided in an embodiment of this application; Figure 3 This is a schematic diagram of the structure of an electronic device provided in an embodiment of this application.

[0018] Explanation of key component symbols: 200. Tunnel Boring Machine Jam Risk Early Warning System; 210. Jam Event Classification Module; 220. Risk Indicator Determination Module; 230. Data Acquisition and Processing Module; 240. First Risk Prediction Module; 250. Second Risk Prediction Module; 260. Risk Early Warning Determination Module; 300. Electronic Equipment; 310. Processor; 320. Communication Interface; 330. Memory; 340. Communication Bus. Detailed Implementation

[0019] To make the objectives, technical solutions, and advantages of this application clearer, the technical solutions of this application will be further described clearly and completely below with reference to the accompanying drawings of the embodiments. It should be noted that the described embodiments are merely some embodiments of this application, and not all embodiments. All other embodiments obtained by those skilled in the art based on the embodiments of this application without creative effort are within the scope of protection of this application.

[0020] Furthermore, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of technical features indicated. Thus, a feature defined as "first" or "second" may explicitly or implicitly include one or more of that feature. In the description of this application, "multiple" means two or more, unless otherwise explicitly specified.

[0021] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application belongs. The terminology used in this specification is for the purpose of describing particular embodiments only and is not intended to be limiting of this application.

[0022] Currently, existing methods for predicting tunnel boring machine (TBM) jamming risks in complex geological environments on plateaus suffer from the following main limitations: First, they lack dynamic analysis of the multi-factor coupling effects under the unique geological conditions of plateaus, failing to systematically consider the synergistic mechanisms of geological hazards such as high ground stress, large deformation of soft rock, and rockbursts, resulting in poor accuracy in risk warning and assessment. Second, existing methods primarily employ offline analysis, exhibiting weak real-time monitoring capabilities and difficulty in timely warning of sudden jamming accidents. Furthermore, most prediction models are based on fixed threshold settings, leading to not only a high false alarm rate but also difficulty adapting to the significant geological variations characteristic of plateau regions. These limitations directly result in insufficient timeliness of warnings and poor assessment accuracy, failing to meet the practical engineering needs for precise quantitative assessment of jamming risks under complex working conditions.

[0023] Example 1 The tunnel boring machine jamming risk warning method provided in this application effectively solves the problem of insufficient accuracy in quantitative assessment of tunnel boring machine jamming risk under complex geological conditions in plateau regions. Figure 1 This is a schematic flowchart of the tunnel boring machine jam risk warning method provided in the embodiments of this application, as shown below. Figure 1 The method includes the following steps: S100. Classify the types of machine jamming events of tunnel boring machines based on historical event data, and determine multiple jamming event types and corresponding preset jamming risk levels.

[0024] As an optional implementation method of this application, historical event data of tunnel boring machine (TBM) jamming incidents that occurred during the construction of tunnels in complex geological environments on plateaus both domestically and internationally can be obtained. Based on the historical event data, the situation and causes of the jamming accidents are analyzed. The jamming event types of TBMs are classified according to geological exploration data, TBM tunneling parameters, and historical case databases, resulting in four types of jamming events: fault fracture zone collapse jamming, alteration zone sudden surge jamming, soft rock large deformation jamming, and hard rock rock burst jamming.

[0025] Fault fracture zone collapse jamming is caused by the development of fissures, such as the instability and collapse of fault breccia under the softening of fissure water and vibration disturbance, blocking the tunneling space. Alteration zone sudden surge jamming is caused by groundwater or weak interlayers, such as the disintegration of chloritized and kaolinized rock masses upon encountering water and the lubrication of interlayers, inducing mudslides and water surges that bury large amounts of mud and sand in the cutterhead and clog the main unit. Soft rock large deformation jamming is caused by the plastic deformation of the surrounding rock, such as the rheological and plastic intrusion of weak surrounding rocks such as claystone and carbonaceous slate under high ground stress, causing the shield to be tightened. Hard rock rockburst jamming is caused by the instantaneous release of high ground stress, such as the sudden release of high ground stress in brittle granite and marble triggering a chain propagation of micro-fractures, causing spalling and jamming of the cutterhead.

[0026] Furthermore, for the four types of lag events mentioned above, a preset lag risk level is assigned to each type of lag event based on relevant data from historical event data.

[0027] For example, in the case of tunnel boring machine (TBM) getting stuck in a fault fracture zone due to a collapse, the risk level can be pre-defined into three levels: minor, moderate, and severe, based on the scale of the collapse and the difficulty of escape from historical event data. Minor stuckness corresponds to no collapse or a small-scale collapse, which can be resolved by adjusting the TBM's excavation parameters, with low difficulty. Moderate stuckness corresponds to a moderate-scale collapse, requiring grouting combined with short-distance widening excavation for escape, with moderate difficulty. Severe stuckness corresponds to a large-scale collapse, requiring a comprehensive approach including grouting, long-distance widening excavation, and advanced support for escape, with high difficulty.

[0028] For example, for sudden water inrush jamming in altered zones, the risk level of the jamming can be pre-defined into three levels: minor, moderate, and severe, based on historical event data regarding water inrush volume, collapse scale, and difficulty of escape. Minor inrush jamming is characterized by dripping water on the surrounding rock surface, with a moderate inrush volume, primarily dripping, with localized linear water flow, no collapse or only small-scale collapse, and low difficulty of escape; it can be resolved by grouting and adjusting the tunnel boring machine's parameters. Moderate inrush jamming is characterized by water flowing through fissures in the surrounding rock, with a large inrush volume, primarily linear, with localized streams of water, accompanied by a moderate-scale collapse, and moderate difficulty of escape; it requires grouting, widening excavation, and backfilling for escape. Severe inrush jamming is characterized by mudflow from the surrounding rock pores, with streams of water carrying mud and sand, accompanied by large-scale collapse, and high difficulty of escape; it requires complex methods such as detours for escape and passage.

[0029] For example, for tunnel boring machines stuck in soft rock with large deformation, the risk level of the stuck machine can be pre-defined into three levels based on the degree of large deformation and the difficulty of unblocking in historical event data: minor large deformation stuck machine, moderate large deformation stuck machine, and severe large deformation stuck machine. Minor large deformation stuck machine is characterized by localized deformation of the surrounding rock. Moderate large deformation stuck machine is characterized by cracking of the support structure, requiring widening excavation. Severe large deformation stuck machine is characterized by tilting of the tunnel boring machine body, requiring widening excavation and reinforcement.

[0030] For example, for tunnel boring machines stuck in hard rock due to rockbursts, the risk level can be pre-defined as either minor or moderate rockburst stuck based on historical event data showing the scale of rockburst collapses and the difficulty of escape. Minor rockburst stuckness manifests as a moderate-scale collapse, which can be resolved by adjusting the tunneling parameters of the tunnel boring machine. Moderate rockburst stuckness manifests as a moderate-scale collapse, requiring measures such as widening the excavation for escape, advanced support to prevent further stuckness, and borehole depressurization.

[0031] Based on this, by classifying jamming event types, a precise mapping between geological risk sources and failure modes can be achieved, improving the targeting of early warnings. For each type of jamming event, a pre-defined jamming risk level can be quantified based on the difficulty of escape and the engineering consequences, allowing for operational response space in the early warning results. The synergy of these two approaches can reduce the false alarm rate of subsequent jamming risk warnings and enhance the accuracy of identifying and the timeliness of handling sudden jamming events in complex plateau geological conditions.

[0032] S200. Identify geological risk sources for the target construction section of the tunnel boring machine and determine the risk indicators and quantitative standards for each type of machine jamming event.

[0033] In this embodiment, the jamming risk sources of the target construction section can be divided into four types based on the four jamming event types of tunnel boring machines: fault fracture zone, weathering and alteration zone, soft rock large deformation section, and hard rock burst section. Geological risk sources are identified by combining tunnel risk assessment and management specifications and geological exploration data, and risk factors of the four types of risk sources are identified. At the same time, the risk index and index quantification standard of each jamming event type are established by using numerical simulation method.

[0034] As an optional implementation method of this application, geological risk source identification can adopt long-distance advanced detection combined with close-range surrounding rock sensing technology. Based on the analysis of geological exploration data, the results of long-distance advanced detection are used as a preliminary judgment, and the results of close-range surrounding rock sensing are used as a close-range accurate prediction to identify the risk source of jamming of the tunnel boring machine.

[0035] For example, targeting risk sources in fault fracture zones, the fault size and characteristics are first identified at a distance using horizontal acoustic profiling or seismic wave reflection. Then, the induced polarization method is used to identify the water-bearing capacity of the fault, compensating for the blind spot of seismic methods regarding water content. Finally, at close range, rock debris image analysis, cutterhead vibration testing, and tunneling parameter analysis are employed. Rock debris image analysis can identify the development of joints and weak interlayers in the surrounding rock in real time; the spectral characteristics of the cutterhead vibration test reflect precursors to face instability; and abnormal combinations of tunneling parameters indicate drastic changes in the interlocking resistance of the surrounding rock. Through these multiple methods, collapse risk prediction shifts from empirical qualitative assessment to data-driven quantitative and qualitative fusion judgment, thereby determining risk indicators for collapse and machine jamming in fault fracture zones. These risk indicators include, but are not limited to, fault size, water-bearing capacity of the fault, degree of fault fracture, inclusions within the fault, equipment functionality, and support system.

[0036] Optionally, by combining the results of on-site risk source identification with historical case research and analysis, quantitative standards for various risk indicators of landslide jamming in fault fracture zones can be established, as shown in Table 1 below.

[0037] Table 1. Schematic diagram of the quantitative standards for risk indicators of landslides in fault fracture zones.

[0038] For example, to address the risk of machine jamming due to sudden inrush in alteration zones, the following methods are employed: first, horizontal acoustic profiling is used at a distance to identify the scale of the alteration zone; then, induced polarization is used to identify the water-bearing capacity of the alteration zone; finally, borehole sampling is used at close range to determine the degree of rock alteration and identify fault locations. The comprehensive test results determine the risk indicators for machine jamming due to sudden inrush in alteration zones. These risk indicators include, but are not limited to, the scale of the alteration zone, the water-bearing capacity of the alteration zone, equipment functionality, and the support system.

[0039] Optionally, the risk indicators in the design and construction phases can be predicted from tunnel design data, induced polarization method and tunnel boring machine parameters, respectively. Based on the model in the geological risk source identification work, the quantitative standards of various risk indicators for sudden jamming in the alteration zone are established as shown in Table 2 below.

[0040] Table 2. Schematic diagram of the quantitative standards for risk indicators of sudden jamming in the alteration zone.

[0041] For example, to address the risk source of large deformation in soft rock, the following steps are taken: First, a comprehensive horizontal acoustic profiling method is used at a long distance to identify weak surrounding rock. Then, design data is analyzed to determine the high ground stress in the section. Finally, surrounding rock parameters are acquired at close range, and machine learning prediction methods are used to identify the degree of fracturing of the weak surrounding rock. The risk indicators for machine jamming due to large deformation in soft rock are determined by combining the results of various analyses. These risk indicators include, but are not limited to, indicators such as the large deformation level, ground stress, rock strength, equipment functionality, and support system.

[0042] Optionally, the risk index values ​​can be determined by combining on-site measurement results with historical case studies. Since the values ​​of risk indicators differ between the design and construction stages, the quantitative standards for various risk indicators of the soft rock large deformation machine can be obtained by inverting from tunnel design data, on-site measurements, and tunneling parameters of the tunnel boring machine, as shown in Table 3.

[0043] Table 3. Schematic diagram of the quantitative standards for risk indicators of large deformation jacking machines in soft rock.

[0044] For example, for the risk source of rockburst in hard rock, the risk source can be identified by combining microseismic monitoring technology, borehole condition analysis and rock sample analysis to obtain the risk index of rockburst jamming in hard rock. The risk index includes, but is not limited to, indicators such as rockburst level, rock strength and ground stress.

[0045] Optionally, the risk indicators for the design and construction phases are obtained by inversion from the tunnel design data and the parameters of the tunnel boring machine, respectively. Based on the model in the risk source identification work, the quantitative standards for various risk indicators of the hard rock blasting machine are established as shown in Table 4 below.

[0046] Table 4. Schematic diagram of the quantitative standards for risk indicators of hard rock blasting machines

[0047] Based on this, by constructing a multi-dimensional dynamic risk indicator system for various card-related event types, the fuzzy and empirical risk source identification is transformed into measurable and traceable quantitative standards, which significantly improves the objectivity and reproducibility of risk assessment and enhances the response accuracy and decision support capability of risk control.

[0048] S300: Acquire multi-source heterogeneous data of the tunnel boring machine, perform data preprocessing on the multi-source heterogeneous data, and obtain cutterhead characteristic data and target characteristic data corresponding to risk indicators.

[0049] In this embodiment, an intelligent sensor network can be deployed on the tunnel boring machine to collect geological environment data and tunneling parameters in real time at the tunnel face. Combined with the aforementioned risk source identification and data collection results, relevant risk indicator data can be obtained, such as fault size, fault water abundance, and alteration zone size. Furthermore, the geological environment data, tunneling parameters, and risk indicator data are integrated to obtain multi-source heterogeneous data.

[0050] As an optional implementation method of this application, during the design and exploration phase, initial macroscopic geological environment data such as the distribution of in-situ stress field, hydrogeological conditions, and surrounding rock classification can be obtained from geological exploration data and design drawings. During the construction phase, a comprehensive technical system combining long-distance advanced detection and close-range surrounding rock sensing can be used to collect real-time geological environment data and risk indicator data in a layered and dynamic manner. For example, induced polarization methods can be used to detect water-rich areas and obtain the water-richness of faults and alteration zones; horizontal acoustic profiling and seismic wave reflection methods can be used to detect the scale and spatial location of faults and alteration zones; and a microseismic monitoring system can be used to monitor micro-fracture events in front of and around the tunnel face in real time, inverting in-situ stress concentration areas and rockburst risk levels.

[0051] Furthermore, geological environmental data can be obtained through close-range sensing by deploying intelligent sensors such as ground-penetrating radar and stress gauges within the tunnel boring machine and tunnel. For example, ground-penetrating radar can monitor the deformation and loosening zone development of the surrounding rock around the tunnel boring machine's shield, while stress gauges can directly measure the pressure of the surrounding rock or shield on the support structure, indirectly reflecting ground stress and surrounding rock deformation. This geological environmental data includes, but is not limited to, groundwater parameters, ground stress parameters, and physical and mechanical parameters of the surrounding rock. For instance, groundwater parameters include, but are not limited to, groundwater pressure, water content, and water abundance, which are key parameters for assessing risks such as inrush and softening interlayers. Ground stress parameters include, but are not limited to, maximum principal stress, minimum principal stress, stress direction, and lateral pressure coefficient, which are core parameters for assessing rockburst risk and large deformation risk in soft rock. Physical and mechanical parameters of the surrounding rock include, but are not limited to, uniaxial compressive strength, elastic modulus, Poisson's ratio, and integrity coefficient of the rock.

[0052] As an optional implementation of this application, a real-time monitoring system for the tunnel boring machine (TBM) in the target construction section can be deployed to obtain historical and real-time tunneling parameters. These parameters are the result of the interaction between the rock strata and the machine, and are a comprehensive reflection of the environmental mechanical conditions. These parameters include, but are not limited to, cutterhead thrust, cutterhead torque, cutterhead penetration, and cutterhead rotation speed. Simultaneously, by combining real-time rock debris image analysis, cutterhead vibration monitoring, and advanced drilling sampling, the strength, integrity, and water content of the surrounding rock can be accurately assessed at close range.

[0053] In this embodiment of the application, multi-source heterogeneous data acquired in real time are preprocessed to obtain target feature data corresponding to cutter head feature data and risk indicators. Data preprocessing includes, but is not limited to, data cleaning, data standardization, feature engineering and correlation analysis.

[0054] As an optional implementation method of this application, data processing tools can be used to clean multi-source heterogeneous data. For example, data integrity checks can be performed on geological environment data and risk indicator data. For missing local parameters in the geological exploration report, interpolation can be used to complete them based on data from previous and subsequent mileage segments, or reasonable filling can be performed based on engineering analogy experience. At the same time, annotations can be added to the dataset. For tunneling parameters, abnormal data such as zero values, negative values, or values ​​far exceeding the equipment's rated values ​​caused by momentary sensor failures and communication interruptions can be identified and removed.

[0055] As an optional implementation of this application, continuous numerical data such as cutterhead thrust, cutterhead torque, in-situ stress, groundwater pressure, and rock strength can be standardized. This standardization includes, but is not limited to, minimum-maximum normalization and Z-score standardization. For geological grade data and ordinal data such as surrounding rock classification (e.g., I, II, III, IV), rockburst intensity, and water abundance (weak, moderate, strong), risk quantification can be performed using an expert scoring system based on the aforementioned risk indicators.

[0056] As a further implementation of the embodiments of this application, feature engineering and correlation analysis can be performed on multi-source heterogeneous data, including but not limited to internal autocorrelation analysis of tunneling parameters, correlation analysis between tunneling parameters and geological environment parameters, and analysis of multi-factor coupling mechanisms and risk indicators.

[0057] For example, internal autocorrelation analysis of tunneling parameters calculates the Pearson correlation coefficient between each pair of tunneling parameters and visualizes it as a heatmap. If two parameters are highly correlated, for example, with a Pearson correlation coefficient > 0.9, then the information they carry is highly overlapping. In this case, the parameter with the clearer physical meaning is retained to eliminate multicollinearity and prevent model overfitting. Correlation analysis between tunneling parameters and geological environment parameters calculates the correlation coefficient between key tunneling parameters and surrounding rock conditions, selecting tunneling parameters that are significantly correlated with changes in the geological environment as core features.

[0058] It is understandable that the causal correlation between risk factors can amplify or inhibit risks, a phenomenon known as the coupling effect of risk factors. For example, when groundwater and weak materials within a fault fracture zone coexist, the groundwater worsens the rock structure and mechanical properties of the fault fracture zone, while the fault fracture zone enhances the weakening effect of groundwater, ultimately leading to machine jamming. Therefore, in this embodiment, a coupling coefficient can be established based on Bayesian conditional probability theory and used for subsequent risk assessment and prediction. For instance, using a tunnel boring machine jamming case as an example, a Bayesian network model of risk factors is established to identify risk factors with coupling relationships. Observation indicators and judgment criteria are defined for each risk factor, forming an extended Bayesian network. Based on this, orthogonal working conditions are designed, and numerical simulation methods are used to calculate the values ​​of the observation indicators of the risk factors. Combined with the judgment criteria, the jamming risk probability is obtained. Finally, the coupling coefficient between the two risk factors is calculated based on the observation indicator values ​​and the jamming risk probability.

[0059] Based on this, by synchronously collecting geological environment data, tunneling parameters and risk index data, the spatiotemporal resolution and working condition coverage of the data are significantly improved. Data preprocessing effectively eliminates the deviations caused by sensor drift, communication delay and modal heterogeneity, providing a reliable data base for subsequent smart card machine risk early warning, and effectively improving the construction safety, efficiency and interpretability of tunnel boring machines.

[0060] S400: Calculate the error rate based on the tool turret feature data and the target tool turret threshold, and determine the risk probability level of each type of jamming event based on the error rate.

[0061] In this embodiment of the application, the risk probability level of each type of SIM card malfunction is determined based on machine learning-based SIM card malfunction risk prediction, specifically including the following steps: First, a tool turret threshold prediction model is constructed, and the tool turret threshold data is predicted based on tool turret feature data and target feature data.

[0062] Understandably, to construct an accurate cutterhead threshold prediction model, it is necessary to select the most representative input variables from massive amounts of data. Therefore, based on the autocorrelation analysis results of the tunneling parameters, the intrinsic correlation between various tunneling parameters can be analyzed using heatmap methods to avoid introducing highly collinear redundant variables. Simultaneously, based on the correlation analysis results between tunneling parameters and surrounding rock, the correlation between key tunneling parameters and surrounding rock conditions can be analyzed by quantifying correlation coefficients, thereby selecting core parameters that are sensitive to changes in surrounding rock and can directly reflect tunneling resistance, ultimately determining them as the input variables for the cutterhead threshold prediction model.

[0063] In this embodiment of the application, the input variables of the cutterhead threshold prediction model are determined to be cutterhead thrust, cutterhead rotation speed, cutterhead torque, cutterhead penetration, surrounding rock strength, surrounding rock integrity, and surrounding rock grade by screening the results of the internal autocorrelation analysis of the tunneling parameters and the correlation analysis results between the tunneling parameters and the surrounding rock.

[0064] As an optional implementation of this application, based on the input variables of the cutterhead threshold prediction model determined above, the total cutterhead thrust and cutterhead torque can be selected as the core prediction output indicators of the cutterhead threshold prediction model. By comprehensively applying three machine learning algorithms—random forest, BP neural network, and support vector machine—a fusion prediction model is constructed that can predict the theoretical cutterhead thrust and cutterhead torque under current geological conditions based on real-time input parameters. After training and parameter enhancement of the fusion prediction model using historical data related to the input variables, the cutterhead threshold prediction model is obtained. Then, based on the cutterhead feature data and target feature data, real-time data related to the input variables is selected and input into the trained cutterhead threshold prediction model for prediction to obtain cutterhead threshold data, which may include the cutterhead thrust threshold and the cutterhead torque threshold.

[0065] Furthermore, based on the cutterhead characteristic data, the real-time cutterhead thrust and real-time cutterhead torque of the tunnel boring machine are selected. A first error rate is calculated based on the real-time cutterhead thrust and cutterhead thrust threshold, and a second error rate is calculated based on the real-time cutterhead torque and cutterhead torque threshold. The first error rate is compared with a first preset error threshold, and the second error rate is compared with a second preset error threshold. When the error rate exceeds the preset error threshold, it indicates that the actual tunneling resistance has increased abnormally and deviated from the normal prediction range, thus determining the risk of machine jamming, thereby achieving early warning of risk. Therefore, the risk probability level can be determined based on the comparison results, and the first and second preset error thresholds can be set according to the actual situation.

[0066] For example, if both the first preset error threshold and the second preset error threshold are set to 20%, the risk probability level includes levels 1-3, with level 3 being the most severe. If the first error rate is greater than or equal to 20% and both the first and second error rates are greater than 20%, the risk probability level is determined to be 3; if either the first or second error rate is greater than 20%, the risk probability level is determined to be 2; and if the first error rate is greater than or equal to 20% and both the first and second error rates are less than or equal to 20%, the risk probability level is determined to be 1.

[0067] Based on this, the input variables obtained through dual correlation screening are fused and modeled to accurately characterize the nonlinear response of rock strata and tunnel boring machines under high ground stress. The risk judgment mechanism driven by the error rate of real-time data and model prediction of cutterhead threshold breaks through the limitation of fixed threshold, effectively advances the early warning response time, and improves the recognition rate of sudden machine jamming.

[0068] S500: Calculate the dynamic risk assessment weight of each risk indicator in each card event type, and determine the card risk level of each card event type based on the dynamic risk assessment weight and target characteristic data.

[0069] In this embodiment of the application, dynamic risk assessment weights are calculated for all risk indicators in each card-operated event type, specifically including the following steps: S501. Determine the best and worst indicators for each risk indicator, and calculate the target subjective weight of each risk indicator based on the best and worst indicators.

[0070] In this embodiment, the subjective weight of the target can be determined based on expert experience and questionnaire surveys using the best-worst method. The core idea is for decision-makers to identify the best and worst extreme indicators among various risk indicators, then compare the relative importance of these two extreme indicators with all other risk indicators, and finally calculate the subjective weight of each risk indicator. The calculation of the target subjective weight specifically includes the following steps: First, determine the set of assessment indicators for each risk indicator, and then identify the best and worst indicators based on this set. Specifically, expert experience and questionnaire surveys can be used to discuss and analyze the risk indicators for each type of SIM card incident to determine the set of assessment indicators. C ={ C 1, C 2, ..., C n},in C i Indicates the first i One risk indicator, i =1,2,…, n , n This indicates the number of risk indicators. Based on expert experience and analysis of actual circumstances, the optimal indicators are identified from the set of assessment indicators. C B and worst indicators C W Among them, the best indicator C B It has a decisive impact on the decision-making outcome, while the worst-case indicator C W The impact is relatively minimal.

[0071] Then, the importance ratio of each evaluation indicator in the optimal indicator and the evaluation indicator set is calculated to obtain the first comparison vector. Specifically, the importance ratio can be calculated based on the importance scores of the optimal indicator and the remaining evaluation indicators, using the following formula: ; In the above formula,a Bi Indicates the best indicator C B Compared to the first i Risk indicators C i The importance ratio can be understood as follows: the first... i Risk indicators C i Excluding best indicators C B The corresponding risk indicators. The importance rating scale can be 1-9, where 1 indicates equal importance, 3 indicates slightly important, 5 indicates significantly important, 7 indicates strongly important, 9 indicates extremely important, and 2, 4, 6 and 8 are intermediate values.

[0072] Calculate the importance ratio of the optimal indicator relative to the other evaluation indicators, and construct a first comparison vector based on the importance ratio. A B =[ a B1 , a B2 ,…, a Bn ].

[0073] Similarly, calculate the importance ratio of the worst-case indicator and each evaluation indicator in the evaluation indicator set to obtain the second comparison vector. Specifically, the importance ratio can be calculated based on the importance scores of the remaining evaluation indicators and the best indicator, using the following formula: ; In the above formula, a iW Indicates the first i Each risk indicator relative to the worst indicator C W The importance ratio can be understood as follows: the first... i Risk indicators C i Excluding worst-case indicators C B The corresponding risk indicators can also be scored using a 1-9 scale.

[0074] Calculate the importance ratio of each of the remaining evaluation indicators relative to the worst indicator, and construct a second comparison vector based on the importance ratio. A W =[ a 1W , a 2W ,…, anW T 。

[0075] Then, construct a subjective weight optimization function based on the best index, the worst index, the first comparison vector, and the second comparison vector. Let the subjective weight vector w = w 1, w 2,…, w n , where w i represents the subjective weight of the i -th risk index, i = 1, 2, …, n , n represents the number of risk indices. Take the difference between the optimal preference score of any i -th risk index and the proportion of the importance degree of the optimal index relative to the i -th risk index to be the smallest, and the difference between the worst preference score of any i -th risk index and the proportion of the importance degree of the i -th risk index relative to the worst index to be the smallest as the optimization objective, and construct the subjective weight optimization function as follows: ； In the above formula, w i represents the subjective weight of the i -th risk index, w B represents the subjective weight of the best index, w W represents the subjective weight of the worst index, a Bi represents the proportion of the importance degree of the optimal index relative to the i ​​​​​​​​​​​​​​​​​​​​​

[0078] Understandably, by combining the above steps with the standards, expert experience, and engineering practice of tunnel boring machine risk assessment, risk indicators for various types of machine jamming events during tunnel boring machine construction are analyzed to determine the best and worst indicators of risk indicators. The best indicators are compared with other risk indicators, and other risk indicators are compared with the worst indicators to obtain the best preference score of the best indicators relative to other risk indicators, and the worst preference score of other risk indicators relative to the worst indicators. The target subjective weight of each risk indicator is then calculated.

[0079] For example, if the best and worst indicators in the risk indicators of landslides in fault fracture zones are the fault fracture zone and groundwater, respectively, then the best preference score of each risk indicator relative to the best indicator, the worst preference score relative to the worst indicator, and the target subjective weight are shown in Table 5 below.

[0080] Table 5. Schematic diagram of the subjective weights of the risk indicators for landslides in fault fracture zones.

[0081] For example, if the best and worst indicators in the risk index of sudden inrush jamming in the altered zone are groundwater and improper selection of tunnel boring machine, respectively, then the best preference score of each risk index relative to the best indicator, the worst preference score relative to the worst indicator, and the target subjective weight of each risk index in the risk index of sudden inrush jamming in the altered zone are shown in Table 6 below.

[0082] Table 6. Schematic diagram of the subjective weights of the risk indicators for sudden jamming in the alteration zone.

[0083] For example, if the best and worst indicators in the risk indicators of large deformation jacking in soft rock are weak surrounding rock and high ground stress, respectively, then the best preference score of each risk indicator relative to the best indicator, the worst preference score relative to the worst indicator, and the target subjective weight of each risk indicator in large deformation jacking in soft rock are shown in Table 7 below.

[0084] Table 7. Schematic diagram of the subjective weights of the risk indicators for large deformation jacking machines in soft rock.

[0085] For example, if the best and worst indicators in the risk indicators of hard rock blasting machine are high ground stress and insufficient monitoring equipment accuracy, respectively, then the best preference score of each risk indicator in the hard rock blasting machine relative to the best indicator, the worst preference score relative to the worst indicator, and the target subjective weight are shown in Table 8 below.

[0086] Table 8. Schematic diagram of the subjective weights of risk indicators for hard rock blasting machines.

[0087] Based on this, by integrating expert experience and multi-source uncertainty information, the rationality and robustness of decision weight allocation can be improved, and the adaptability and interpretability of risk warning methods in fuzzy and small sample scenarios can be enhanced.

[0088] S502. Calculate the index entropy value based on the risk index data of each risk index in historical card-related events, and determine the target objective weight of each risk index based on the index entropy value.

[0089] In this embodiment, objective weights can be determined by introducing statistical data from historical card-related events and other methods devoid of subjective intent. Based on objective historical statistical data, objective weights are determined by calculating the entropy values ​​of each risk indicator. That is, the lower the information entropy of a risk indicator, the greater the amount of information contained in the data; the more significant the differences in risk indicator data, the higher the objective weight. The calculation process of the target objective weight specifically includes the following steps: First, obtain risk indicator data for each risk indicator under multiple historical system crashes. Then, construct an indicator data matrix based on each risk indicator and its data. Assume there are... n Each risk indicator, to obtain m Risk indicator data from previous historical card-related incidents, used to construct an indicator data matrix, is as follows: ; In the above formula, X Represents the indicator data matrix, x ji Indicates the first j The first historical card jam incident i Risk indicator data for each risk indicator, among which j=1,2,…,m , i=1,2,…,n .

[0090] The indicator data matrix is ​​then standardized to obtain a standardized data matrix. Standardization eliminates the dimensions of risk indicators. For risk indicators where construction risk decreases as parameter values ​​increase, the standardization formula is as follows: ; In the above formula, x ji Indicates the first j The first historical card jam incident i Standardized data for each risk indicator, min x j Indicates the first j The smallest risk indicator data in the previous historical card machine incident, max x j Indicates the first jThe most significant risk indicator data in this historical card-related incident.

[0091] The standardized formula for the risk index, which increases construction risk as the parameter value increases, is as follows: ; The standardized data matrix obtained from the calculation is as follows: ; In the above formula, R Represents the indicator data matrix, r ji Indicates the first j The first historical card jam incident i Standardized data for each risk indicator.

[0092] Next, normalize each element of the standardized data matrix to obtain a standard normalized data matrix. The formula for normalizing each element of the standardized data matrix is ​​as follows: ; In the above formula, p ji Indicates the first j The first historical card jam incident i Standardized normalized data for each risk indicator.

[0093] The resulting standard normalized data matrix is ​​as follows: ; In the above formula, P This represents a standard normalized data matrix.

[0094] Finally, the entropy value of each risk indicator is calculated based on the standard normalized data matrix, and the objective weight is calculated based on the indicator entropy value to obtain the target objective weight. i The formula for calculating the entropy value of each risk indicator is as follows: ; In the above formula, e i Indicates the first i The entropy value of each risk indicator.

[0095] Based on the i The objective weight is calculated based on the entropy value of the risk indicator to obtain the first risk indicator. i The objective weights of each risk indicator are calculated using the following formula: ; In the above formula, K i Indicates the first i The objective weight of each risk indicator.

[0096] For example, combining the risk indicator data of historical jamming events in 400 sections of 10 tunnels collected, the above-mentioned methods and steps are used to calculate the index entropy value and objective weight of each risk indicator, extract the objective weight of each risk indicator for the four types of jamming events, and perform normalization processing to obtain the target objective weight of each risk indicator for the four types of jamming events. The target objective weights of each risk indicator in the fault fracture zone collapse jamming are shown in Table 9 below; the target objective weights of each risk indicator in the alteration zone surge jamming are shown in Table 10 below; the target objective weights of each risk indicator in the soft rock large deformation jamming are shown in Table 11 below; and the target objective and subjective weights of each risk indicator in the hard rock rockburst jamming are shown in Table 12 below.

[0097] Table 9. Schematic diagram of the objective weights of risk indicators for landslides and malfunctions in fault fracture zones.

[0098] Table 10. Target Objective Weights of Risk Indicators for Sudden Machine Jams in Alteration Zones

[0099] Table 11. Schematic diagram of objective weights for risk indicators of large deformation jacking machines in soft rock.

[0100] Table 12. Schematic diagram of objective weights for risk indicators of hard rock blasting machines.

[0101] Based on this, the calculation of objective weights is automatically weighted by historical risk indicator data, avoiding subjective bias, improving the repeatability and credibility of evaluation results, and taking into account the conflict between indicators and the amount of information, so that the weight allocation is more in line with the actual data distribution and enhances the robustness and interpretability of the decision-making model.

[0102] S503. Assign weights to the subjective and objective weights of each risk indicator to obtain the combined weights of each risk indicator.

[0103] In this embodiment, the subjective and objective weights of the target can be combined, taking into account both expert experience and objective data characteristics, to obtain a combined weight that is more reasonable. Optionally, a game theory-based combined weighting method can be used to combine the subjective and objective weights of the target, specifically including the following steps: If it exists n Each risk indicator has a target subjective weight vector. w =( w 1, w 2,…, wn ), Target objective weight vector K =( K 1, K 2,…, K n If so, the combined weight vector after the combination is completed is set as follows: ; In the above formula, Z represents the combined weight vector. The first combination coefficient represents the subjective weight of the objective. The second combination coefficient represents the objective weight of the target. The column vector form representing the target subjective weight vector. This represents the column vector form of the objective weight vector.

[0104] Then, to optimize the combination coefficients, based on game theory, we need to find a set of combination coefficients. This combination of coefficients The optimal linear combination ensures that the resulting combined weight vector Z is similar to the target subjective weight vector. w and target objective weight vector K The overall deviation is minimized. Based on this, the combination coefficient function is constructed as follows: ; It is understandable that solving the above function is an unconstrained optimization problem, which can be applied to... Find the first-order partial derivatives, and set them equal to 0. Then, find the combination coefficient vector that corresponds to the minimum value of the combination coefficient function. The formula for calculating the first derivative condition of the linear equation system is as follows: ; The above system of linear equations is actually a system of two linear equations in two variables. By solving this system of two linear equations, we can obtain the combination coefficient vector. In and The optimal value is then standardized to obtain the final combination coefficient vector expression. as follows: ; In the above formula, This represents the normalized optimal first combination coefficient of the objective's subjective weights. This represents the optimal second combination coefficient after normalization of the objective weights of the target.

[0105] Finally, substituting the normalized optimal first combination coefficient and optimal second combination coefficient into the linear combination formula, we obtain the combination weights for each risk indicator as follows: ; In the above formula, Z i Indicates the first i The combined weights of individual risk indicators w i Indicates the first i The target subjective weight of each risk indicator K i Indicates the first i The objective weight of each risk indicator.

[0106] Based on this, the final combined weight vector Z={ Z 1, Z 2, ..., Z n}satisfy .

[0107] For example, based on the target subjective weight and target objective weight of each risk indicator for the four types of jamming events mentioned above, a game theory-based combined weighting method is used to calculate the combined weight of each risk indicator for the four types of jamming events. The combined weights of each risk indicator in the fault fracture zone collapse jamming event are shown in Table 13 below; the combined weights of each risk indicator in the alteration zone sudden surge jamming event are shown in Table 14 below; the combined weights of each risk indicator in the soft rock large deformation jamming event are shown in Table 15 below; and the combined weights of each risk indicator in the hard rock rockburst jamming event are shown in Table 16 below.

[0108] Table 13. Schematic diagram of the combined weights of risk indicators for landslides in fault fracture zones.

[0109] Table 14. Schematic diagram of the combined weights of risk indicators for sudden jamming in the alteration zone.

[0110] Table 15. Schematic diagram of the combined weights of risk indicators for large deformation jacking machines in soft rock.

[0111] Table 16. Schematic diagram of the combined weights of risk indicators for hard rock blasting machines.

[0112] Based on this, game theory combinatorial weighting treats the subjective weight and objective weight of the target as participants in the game. By coordinating the conflict between the two weights, a combinatorial weight that can be accepted by all participants is found. This comprehensively utilizes subjective and objective information, overcomes the limitations of the single weighting method, significantly improves robustness and discrimination accuracy, avoids the accumulation of bias caused by a single weight, enhances the ability to suppress noise and outliers, and has both interpretability and generalization.

[0113] S504. Based on the contribution scores of each risk indicator to the card machine risk, the combined weights are dynamically adjusted to obtain the dynamic risk assessment weights of each risk indicator.

[0114] In this embodiment of the application, a variable weight vector is constructed based on the contribution scores of each risk indicator to the card machine risk, and the combined weights are dynamically adjusted. Specifically, the steps include the following: First, construct the indicator state vector based on the contribution scores of each risk indicator, and then determine the balance function of the indicator state vector.

[0115] In this embodiment of the application, the indicator state vector is constructed as follows: ; In the above formula, S i ( X ) indicates the first i The indicator state variable value of the first risk indicator reflects the first... i The impact of the current state of each risk indicator on the overall equilibrium. This represents the state variable balance function, used to measure the overall degree of equilibrium of various risk indicator states. x i Indicates the first i The contribution score of each risk indicator to the risk of card malfunction.

[0116] As an optional implementation of this application, the state variable balance function can be a product-type state balance function, and the expression of the state variable balance function is as follows: ; In the above formula, This represents the equilibrium coefficient.

[0117] Then, the combined weights of each risk indicator and the normalized product of the indicator state vectors are calculated to obtain the dynamic risk assessment weights. The calculation formula is as follows: ; In the above formula, W i ( X ) indicates the first i Dynamic risk assessment weights for each risk indicator. Z i Indicates the first i The combined weighting of individual risk indicators. Balance coefficient. This reflects the dynamic nature of risk indicators, when When the value is less than 0.5, the risk indicator weights exhibit weak dynamics, changing only slightly while maintaining the overall portfolio weights and proportions. When the value is ≥0.5, the risk indicator weights are highly dynamic, and there is a possibility of significant changes that disrupt the overall weighting and proportion of the portfolio. Optionally, this could be... Set to 0.5.

[0118] Based on this, in the initial stage, dynamic risk assessment weights are obtained by combining the combined vector and the indicator state vector, where the indicator state vector reflects the real-time status of the evaluated object. Subsequently, the dynamic risk assessment weights can be updated in real time based on the indicator state vector. Once the state of the risk indicator changes, the risk indicator state variable can be recalculated, thereby changing the weight vector and ultimately adjusting the dynamic risk assessment weights to adapt to the dynamic changes in indicator risk.

[0119] In this embodiment, the basic data of risk indicators often exhibit uncertainty. Specifically, the geological environment data and tunneling parameters for a given target construction section are typically distributed in intervals, making them impossible to accurately describe with a single numerical value. Therefore, an improved attribute interval method can be used to integrate the target feature data intervals of multiple risk indicators into a comprehensive jamming risk assessment result through interval calculations, aggregation, and comparisons. This ultimately determines the jamming risk level for each jamming event type. The specific steps for determining the jamming risk level for each jamming event type based on the improved attribute interval method are as follows: S510. Based on risk indicators and preset card machine risk levels, construct minimum value standard matrix and maximum value standard matrix within preset value ranges.

[0120] Assume a certain type of card-related event exists. n Individual risk indicators and d If there are preset card risk levels, then any number of card risk levels will be... i Each risk indicator falls within a preset value range ( a ik , b ik At that time, the first k The attribute measure for the preset card machine risk level is set to 1, where... a ik Indicates the first i Each risk indicator is at the preset card machine risk level. k Minimum risk indicator data at level 1, b ik Indicates the first i Each risk indicator is at the preset card machine risk level. k The maximum risk indicator data at level 1. i= 1,2,…, n , k =1,2,…, d The minimum value standard matrix A and the maximum value standard matrix B are constructed using the minimum and maximum value endpoints of the preset value intervals, respectively, as follows: ; ; in, a ik < b ik And satisfy a i1 < a i2 <...< a id , b i1 < b i2 <...< b id .

[0121] For example, based on attribute measure theory, the quantitative standards for the four types of card-related events are transformed into maximum and minimum standard matrices, respectively. Specifically, the levels of qualitative risk indicators are assigned quantitative intervals, with levels state0 to state3 sequentially quantified as 0-25, 25-50, 50-75, and 75-100. The minimum and maximum standard vectors for this type of risk indicator are respectively... and For quantitative risk indicators, the endpoints of their quantification standard intervals are used to form minimum and maximum standardized vectors, which in turn form minimum and maximum standard matrices. The maximum and minimum standard matrices for the four types of card-related events are shown in Table 17 below.

[0122] Table 17. Schematic diagram of the standard matrix of maximum and minimum values ​​for four types of lag events.

[0123] S520. Calculate multiple single-index attribute measurement matrices based on the target feature data, the minimum value standard matrix, and the maximum value standard matrix.

[0124] In this embodiment of the application, the actual risk indicator data of each risk indicator can be obtained based on the target feature data. Assuming any... i The actual risk indicator data for each risk indicator is: t i =[ t ix , t iy ],in t ix Indicator data of actual risk indicators t i The actual minimum value, tiy Indicator data of actual risk indicators t i The actual maximum value. Based on actual risk indicator data. t i Substituting the endpoints of the interval into the standard matrix A of the minimum value and the standard matrix B of the maximum value, the calculation is performed using the following rules: When actual risk indicator data t i ≤ a i1 When the interval endpoints correspond to the attribute measure, the following rules are defined: ; When actual risk indicator data t i ≤ b i1 When the interval endpoints correspond to the attribute measure, the following rules are defined: ; When actual risk indicator data t i ≥ a id When the interval endpoints correspond to the attribute measure, the following rules are defined: ; When actual risk indicator data t i ≥ b id When the interval endpoints correspond to the attribute measure, the following rules are defined: ; when a ik ≤ Actual risk indicator data t i ≤ a i(k+1) When the interval endpoints correspond to the attribute measure, the following rules are defined: ; when b ik ≤ Actual risk indicator data t i ≤ b i(k+1) When the interval endpoints correspond to the attribute measure, the following rules are defined: ; In the above formulas, Indicator data of actual risk indicators t iThe preset card machine risk level is k The lower limit of the attribute measure at level 1. Indicator data of actual risk indicators t i The preset card machine risk level is k The upper limit of attribute measurement at level 1 i= 1,2,…, n , k =1,2,…, d .

[0125] Based on the above rule definitions, substituting the endpoints of each risk indicator's interval into the minimum value standard matrix A and the maximum value standard matrix B yields four single-indicator attribute measurement matrices. Specifically, substituting the minimum value of each risk indicator's interval into the minimum value standard matrix A yields the first single-indicator attribute measurement matrix. The second single-indicator attribute measurement matrix is ​​obtained by substituting the maximum value of each risk indicator interval into the minimum value standard matrix A. The third single-indicator attribute measurement matrix is ​​obtained by substituting the minimum value of each risk indicator interval into the maximum value standard matrix B. Substituting the maximum value of each risk indicator interval into the maximum value standard matrix B, we obtain the fourth single-indicator attribute measurement matrix. The first single-index attribute measurement matrix It is expressed as follows: ; In the above formula, Indicates the first i Each risk indicator is at the preset card machine risk level. k The lower bound attribute measure of level is substituted into the value of the minimum standard matrix A.

[0126] Second Single-Indicator Attribute Measurement Matrix It is expressed as follows: ; In the above formula, Indicates the first i Each risk indicator is at the preset card machine risk level. k The upper limit attribute measure of the level is substituted into the value of the minimum standard matrix A.

[0127] Third Single-Indicator Attribute Measurement Matrix It is expressed as follows: ; In the above formula, Indicates the first i Each risk indicator is at the preset card machine risk level. k The lower bound attribute measure of the level is substituted into the value of the standard matrix B of the maximum value.

[0128] Fourth Single-Indicator Attribute Measurement Matrix It is expressed as follows: ; In the above formula, Indicates the first i Each risk indicator is at the preset card machine risk level. k The upper limit attribute measure of the level is substituted into the value of the standard matrix B of the maximum value.

[0129] In this embodiment of the application, based on the corresponding first of the four single-index attribute measurement matrices i The row vector of the risk indicator is quantified, and the first risk indicator is calculated based on the quantized value. i The contribution score of each risk indicator to the risk of card malfunction. x i Then the first i The contribution score of each risk indicator to the risk of card malfunction. x i Substitute the dynamic risk assessment weights into the weights for adjustment.

[0130] For example, based on the tunnel engineering data of a certain target construction section, the tunnel boring machine's tunneling section 949+574~934+913 is initially divided into 9 tunneling intervals using the type of adverse geological body as a risk indicator. The interval mileage, interval length, and typical adverse geological bodies of each tunneling interval are shown in Table 18.

[0131] Table 18. Schematic diagram of the section mileage, section length, and typical adverse geological features of the nine tunneling sections.

[0132] By acquiring tunnel construction data, tunnel design data, on-site and indoor mechanical test results, and actual on-site construction data, the risk index data or risk index data ranges for nine tunneling sections were obtained, as shown in Table 19.

[0133] Table 19. Risk Indicator Data or Risk Indicator Data Ranges for 9 Tunneling Sections

[0134] The risk index data or risk index data intervals of the nine tunneling sections are substituted into the maximum value standard matrix and the minimum value standard matrix, respectively. Due to the large number of functions and similar processes, this embodiment only shows the calculation process of the second tunneling section. Assuming that the jamming event type of the second tunneling section is a fault fracture zone collapse jamming, the four single-index attribute measurement matrices are shown in Table 20.

[0135] Table 20. Schematic diagram of the four single-index attribute measurement matrices for landslide jacking machines in fault fracture zones.

[0136] Assuming the jamming event type in the second tunneling section is a sudden jamming in the erosion zone, the four single-index attribute measurement matrices are shown in Table 21.

[0137] Table 21. Schematic diagram of the four single-index attribute measurement matrices for sudden card jamming in the alteration zone.

[0138] Assuming the jamming event type in the second tunneling section is jamming due to large deformation in soft rock, the four single-index attribute measurement matrices are shown in Table 22.

[0139] Table 22. Schematic diagram of the four single-index attribute measurement matrices for large deformation jacks in soft rock.

[0140] Assuming the jamming event type in the second tunneling section is hard rock rockburst jamming, the four single-index attribute measurement matrices are shown in Table 23.

[0141] Table 23. Schematic diagram of the four single-index attribute measurement matrices for hard rock rockburst machines.

[0142] S530. Based on the risk coupling coefficient between each of the risk indicators, the single indicator attribute measurement vector is modified to obtain each modified single indicator attribute measurement vector.

[0143] It is understandable that the simultaneous occurrence of risk factors with coupling effects significantly increases the risk of tunnel boring machine jamming compared to the case where each risk factor is independent. Therefore, the coupling coefficient of the risk factors calculated in step S300 is introduced to correct the measurement values ​​of risk index attributes with coupling effects, in order to more accurately predict the risk of tunnel boring machine jamming under the presence of coupling effects.

[0144] For example, suppose the first p The first risk indicator and the first q There is a coupling effect among the risk indicators, with a coupling coefficient of . , No. p The first risk indicator and the first q The single-attribute measure vectors corresponding to each risk indicator are as follows: and When both exist simultaneously, with (1- ) for the first p The first risk indicator and the first q The single-attribute measure vector of each risk indicator is corrected as follows: ; Then the corrected attribute measure matrix and Each row vector is substituted back and replaced. , , and The corresponding row vectors in the middle are used to obtain the measure vectors of each modified single indicator attribute. , , and .

[0145] For example, in the second tunneling section, the risk index row measurement vector with coupling effect is corrected to obtain the four corrected single index attribute measurement matrices of the fault fracture zone collapse jamming machine as shown in Table 24.

[0146] Table 24. Schematic diagram of the four modified single-index attribute measurement matrices for landslide jacking machines in fault fracture zones.

[0147] Similarly, the four modified single-index attribute measurement matrices for alteration zone sudden surge jamming, soft rock large deformation jamming, and hard rock rockburst jamming can also be obtained through coupling coefficient correction, which will not be shown one by one here.

[0148] S540. The modified single-indicator attribute measurement matrices are fused with the dynamic risk assessment weights to obtain the corresponding multi-indicator attribute measurement vectors.

[0149] In this embodiment, the four single-indicator attribute measurement matrices are first fused with combined weights, and then dynamic risk assessment weights are dynamically generated based on subsequent real-time target feature data for data fusion, resulting in four multi-indicator attribute measurement vectors as follows: ; In the above formula, This indicates that the preset card machine risk level is... k The lower bound attribute measure of level is substituted into the corrected fusion value of the minimum standard matrix A. This indicates that the preset card machine risk level is... k The upper limit attribute measure vector value of the level is substituted into the corrected fusion value of the minimum standard matrix A. This indicates that the preset card machine risk level is... k The lower bound attribute measure of level is substituted into the corrected fusion value of the maximum standard matrix B. This indicates that the preset card reader risk level is... k The upper limit attribute measure of the level is substituted into the corrected fusion value of the standard matrix B with the maximum value.

[0150] S550. Calculate the mean of each multi-index attribute measure vector to obtain the target attribute measure vector.

[0151] As an optional implementation of this application, the mean method can be used to unify the four multi-index attribute measure vectors into a target attribute measure vector. U The calculation formula is as follows: ; In the above formula, u k Indicates the first k The mean of the vector of multiple indicator attributes for the preset risk level of the card machine. U This represents the target attribute measure vector.

[0152] S560. Calculate the risk quantification value based on the target attribute measurement vector and the maximum value of the endpoints of each measurement interval, and determine the lag risk level of each lag event type based on the risk quantification value.

[0153] In this application embodiment, it is assumed that there exists H For each measurement interval, then for H The measurement intervals are quantized, with the first interval starting from 0, and each interval having a width of 100. H Then any number h The minimum and maximum values ​​at the endpoints of each measurement interval are as follows: ; ; In the above formula, L hmin Indicates the first h The minimum value at the endpoints of each measurement interval. L hmax Indicates the first h Maximum value at the endpoints of each measurement interval This process is repeated until the endpoint of the measurement interval reaches 100. The final risk quantification value is obtained by multiplying the maximum value at each endpoint of the measurement interval by the interval attribute measure in the target attribute measurement vector and summing the results: ; In the above formula, C This represents the quantified risk value. Represents the first element in the target attribute measure vector. h Mean of a multi-index attribute measure vector for a given measurement interval.

[0154] By comparing the final risk quantification value with the preset risk quantification threshold, the SIM card risk level of each SIM card event type can be determined. This SIM card risk level is one of the preset SIM card risk levels determined in step S100.

[0155] For example, based on the four multi-index attribute measurement vectors of each of the four jamming event types in the second tunneling section, the risk quantification value of the four jamming event types is calculated, thereby determining the jamming risk level of the four jamming event types as shown in Tables 25-28.

[0156] Table 25. Schematic diagram of the jamming risk level results for landslide jamming in fault fracture zones.

[0157] Table 26. Schematic diagram of jamming risk level results for sudden jamming in the alteration zone.

[0158] Table 27. Schematic diagram of jamming risk level results for large deformation jamming machines in soft rock.

[0159] Table 28. Schematic diagram of jamming risk level results for hard rock rockburst machines.

[0160] Based on this, by dynamically evaluating weight calculations, taking into account both expert experience and data objectivity, and combining a variable weight mechanism to dynamically respond to geological changes, and then using an improved attribute interval model to integrate uncertain data and coupling effect correction, the risk level determination of the card machine is made more robust, interpretable and adaptive, especially suitable for dynamic high-risk scenarios with complex geology on plateaus.

[0161] S600 determines the risk warning level for each type of malfunction event based on the risk probability level and the malfunction risk level.

[0162] In this embodiment of the application, the risk warning level of the corresponding card-operated event type can be determined by comprehensively considering the risk probability level and the card-operated risk level of each event type. The specific determination rules can be set according to the actual situation.

[0163] For example, for a specific construction section, the risk probability levels for each type of machine jamming event are divided into three levels: 1-3, with level 3 being the most severe. The machine jamming risk levels are also divided into three states: State 0-State 2, where State 0 represents a minor jamming, State 1 represents a moderate jamming, and State 2 represents a severe jamming. The final comprehensive risk warning level is determined into four levels: low, medium, high, and very high. The rules for determining the machine jamming risk warning level for a specific construction section are shown in Table 29 below.

[0164] Table 29. Schematic Diagram for Determining the Risk Warning Level of Tunnel Boring Machine Jamming in a Target Construction Section

[0165] As shown in Table 29, if the probability level of the fault fracture zone collapse jamming is 1 and the jamming risk level is State 1, then the risk warning level for the fault fracture zone collapse jamming is determined to be low. If the probability level of the alteration zone surge jamming is 2 and the jamming risk level is State 1, then the risk warning level for the alteration zone surge jamming is determined to be medium. If the probability level of the soft rock large deformation jamming is 2 and the jamming risk level is State 2, then the risk warning level for the soft rock large deformation jamming is determined to be high. If the probability level of the hard rock rockburst jamming is 3 and the jamming risk level is State 2, then the risk warning level for the hard rock rockburst jamming is determined to be extremely high.

[0166] As a further implementation of the embodiments of this application, the construction process can be adjusted in real time by referring to the risk warning level. For example, when the risk warning level is high, an alarm can be set in real time and control measures such as reducing the tunneling speed can be recommended.

[0167] The tunnel boring machine (TBM) jamming risk early warning method provided in this application constructs a refined classification system for jamming event types, enabling early warnings to accurately match risk scenarios with different geological causes, significantly improving the targeting and engineering adaptability of the early warnings. Simultaneously, it introduces a multi-risk factor coupling mechanism analysis, breaking through the linear limitations of traditional single-parameter thresholds and enhancing the predictability of sudden and complex jamming events. Combining a dynamic weighting mechanism and an improved attribute interval model, the risk early warning indicators can adaptively adjust with the real-time evolution of the surrounding rock condition, comprehensively improving the robustness and decision-making reliability of the TBM early warning system under complex and variable geological conditions in plateau regions.

[0168] Example 2 Based on the same technical concept as Embodiment 1 above, this application also provides a tunnel boring machine jamming risk warning system. Figure 2 This is a schematic diagram of the tunnel boring machine jam risk warning system provided in the embodiments of this application, as shown below. Figure 2 As shown, the tunnel boring machine jam risk early warning system 200 includes: The jamming event classification module 210 is used to classify the jamming event types of tunnel boring machines based on historical event data, and to determine multiple jamming event types and corresponding preset jamming risk levels.

[0169] The risk indicator determination module 220 is used to identify geological risk sources in the target construction section of the tunnel boring machine and determine the risk indicators and quantitative standards for each type of machine jamming event.

[0170] The data acquisition and processing module 230 is used to acquire multi-source heterogeneous data of the tunnel boring machine, perform data preprocessing on the multi-source heterogeneous data, and obtain cutterhead feature data and target feature data corresponding to risk indicators.

[0171] The first risk prediction module 240 is used to calculate the error rate based on the tool head feature data and the target tool head threshold, and to determine the risk probability level of each type of machine jamming event based on the error rate.

[0172] The second risk prediction module 250 is used to calculate the dynamic risk assessment weight of each risk indicator in each card event type, and determine the card risk level of each card event type based on the dynamic risk assessment weight and target characteristic data.

[0173] The risk warning determination module 260 is used to jointly determine the risk warning level of each type of card-related event based on the risk probability level and the card-related risk level.

[0174] The tunnel boring machine jam risk early warning system provided in this application embodiment relies on an intelligent sensor network and a hierarchical perception architecture to achieve closed-loop data support from long-distance macroscopic prediction to short-distance dynamic response, ensuring the timeliness and on-site verifiability of risk assessment.

[0175] It is understood that the implementation method of the tunnel boring machine jamming risk warning method in Embodiment 1 above is also applicable to this embodiment and can achieve the same technical effect, so it will not be described again here.

[0176] Example 3 Based on the same concept, this application also provides an electronic device. Figure 3 This is a schematic diagram of the structure of an electronic device provided in an embodiment of this application, such as... Figure 3 As shown, the electronic device 300 may include a processor 310, a communication interface 320, a memory 330, and a communication bus 340, wherein the processor 310, the communication interface 320, and the memory 330 communicate with each other via the communication bus 340. The processor 310 can call logical instructions in the memory 330 to execute the steps of the tunnel boring machine jam risk warning method as described in the above embodiments. For example, it includes: S100. Classify the types of jamming events of tunnel boring machines based on historical event data, and determine multiple jamming event types and corresponding preset jamming risk levels; S200. Identify geological risk sources for the target construction section of the tunnel boring machine and determine the risk indicators and quantitative standards for each type of machine jamming event. S300: Acquire multi-source heterogeneous data of the tunnel boring machine, perform data preprocessing on the multi-source heterogeneous data, and obtain cutterhead feature data and target feature data corresponding to risk indicators; S400: Calculate the error rate based on the tool head feature data and the target tool head threshold, and determine the risk probability level of each type of jamming event based on the error rate; S500: Calculate the dynamic risk assessment weight of each risk indicator in each card event type, and determine the card risk level of each card event type based on the dynamic risk assessment weight and target characteristic data. S600 determines the risk warning level for each type of malfunction event based on the risk probability level and the malfunction risk level.

[0177] The processor 310 can be a central processing unit (CPU). The processor can also be other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, or combinations of the above types of chips.

[0178] Furthermore, the logical instructions in the aforementioned memory 330 can be implemented as software functional units and, when sold or used as independent products, can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or a part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0179] The memory 330 may include a program storage area and a data storage area. The program storage area may store the operating system and applications required for at least one function; the data storage area may store data created by the processor, etc. Furthermore, the memory may include high-speed random access memory and non-transitory memory, such as at least one disk storage device, flash memory device, or other non-transitory solid-state storage device. In some embodiments, the memory may optionally include memory remotely located relative to the processor, which can be connected to the processor via a network. Examples of such networks include, but are not limited to, the Internet, corporate intranets, local area networks, mobile communication networks, and combinations thereof.

[0180] Example 4 Based on the same concept, embodiments of this application also provide a computer-readable storage medium storing a computer program containing at least one piece of code executable by a master control device to control the master control device to implement the steps of the tunnel boring machine jam risk warning method as described in the above embodiments. For example, it includes: S100. Classify the types of jamming events of tunnel boring machines based on historical event data, and determine multiple jamming event types and corresponding preset jamming risk levels; S200. Identify geological risk sources for the target construction section of the tunnel boring machine and determine the risk indicators and quantitative standards for each type of machine jamming event. S300: Acquire multi-source heterogeneous data of the tunnel boring machine, perform data preprocessing on the multi-source heterogeneous data, and obtain cutterhead feature data and target feature data corresponding to risk indicators; S400: Calculate the error rate based on the tool head feature data and the target tool head threshold, and determine the risk probability level of each type of jamming event based on the error rate; S500: Calculate the dynamic risk assessment weight of each risk indicator in each card event type, and determine the card risk level of each card event type based on the dynamic risk assessment weight and target characteristic data. S600 determines the risk warning level for each type of malfunction event based on the risk probability level and the malfunction risk level.

[0181] Based on the same technical concept, this application also provides a computer program, which, when executed by a main control device, is used to implement the above-described method embodiments.

[0182] The computer program may be stored, in whole or in part, on a computer-readable storage medium packaged with the processor, or in part or in whole on a memory not packaged with the processor.

[0183] Based on the same technical concept, embodiments of this application also provide a processor for implementing the above-described method embodiments. The processor may be a chip.

[0184] In summary, the tunnel boring machine (TBM) jamming risk early warning method, system, equipment, and storage medium provided in this application construct a refined classification system for jamming event types, enabling early warnings to accurately match risk scenarios with different geological causes, significantly improving the targeting and engineering adaptability of early warnings. Simultaneously, the introduction of multi-risk factor coupling mechanism analysis overcomes the linear limitations of traditional single-parameter thresholds, enhancing the predictive ability for sudden and complex jamming events. Relying on intelligent sensor networks and a hierarchical perception architecture, closed-loop data support is achieved from long-distance macroscopic prediction to short-distance dynamic response, ensuring the timeliness and on-site verifiability of risk assessment. Combining a dynamic weighting mechanism and an improved attribute interval model, the risk early warning indicators can adaptively adjust with the real-time evolution of the surrounding rock condition, comprehensively improving the robustness and decision-making reliability of the TBM early warning system under complex and variable geological conditions in plateau regions.

[0185] In this document, reference to "embodiment" means that a particular feature, structure, or characteristic described in connection with an embodiment may be included in at least one embodiment of this application. It will be explicitly and implicitly understood by those skilled in the art that the embodiments described herein can be combined with other embodiments.

[0186] The above embodiments are merely illustrative of several implementation methods of this application, and their descriptions are relatively specific and detailed. However, they should not be construed as limiting the scope of this application's patent. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of this application, and these all fall within the protection scope of this application.

[0187] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit them. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of this application.

Claims

1. A method for early warning of machine jamming risk in a tunnel boring machine, characterized in that, The method includes: Based on historical event data, tunnel boring machine jamming event types are classified to determine multiple jamming event types and corresponding preset jamming risk levels; Geological risk sources are identified for the target construction section of the tunnel boring machine, and risk indicators and quantitative standards for each type of machine jamming event are determined. Acquire multi-source heterogeneous data of the tunnel boring machine, perform data preprocessing on the multi-source heterogeneous data, and obtain cutterhead feature data and target feature data corresponding to the risk indicators; The error rate is calculated based on the cutter head feature data and the target cutter head threshold, and the risk probability level of each of the jamming event types is determined based on the error rate. Calculate the dynamic risk assessment weight of each risk indicator in each of the card-operated event types, and determine the card-operated risk level of each of the card-operated event types based on the dynamic risk assessment weight and the target feature data; The risk warning level for each type of card-related event is determined by combining the risk probability level and the card-related risk level.

2. The method for early warning of machine jamming risk of tunnel boring machine according to claim 1, characterized in that, The calculation of the dynamic risk assessment weights for each risk indicator in each of the card-operated event types includes: Determine the best and worst indicators for each of the aforementioned risk indicators, and calculate the target subjective weight for each of the aforementioned risk indicators based on the best and worst indicators; Calculate the index entropy value based on the risk index data of each risk index in historical card events, and determine the target objective weight of each risk index based on the index entropy value; The target subjective weight and target objective weight of each risk indicator are weighted and combined to obtain the combined weight of each risk indicator; The combined weights are dynamically adjusted based on the contribution scores of each risk indicator to the card machine risk, thereby obtaining the dynamic risk assessment weights of each risk indicator.

3. The method for early warning of machine jamming risk of tunnel boring machine according to claim 2, characterized in that, The step of determining the card-operated risk level for each of the card-operated event types based on the dynamic risk assessment weights and the target feature data includes: Based on the risk indicators and the preset card machine risk level, construct a minimum value standard matrix and a maximum value standard matrix within a preset value range; Calculate multiple single-index attribute measurement matrices based on the target feature data, the minimum value standard matrix, and the maximum value standard matrix; The individual single-indicator attribute measurement vectors are modified based on the risk coupling coefficients between the individual risk indicators to obtain the modified individual single-indicator attribute measurement vectors. Each of the modified single-indicator attribute measurement matrices is fused with the dynamic risk assessment weight to obtain the corresponding multi-indicator attribute measurement vector. The target attribute measure vector is obtained by calculating the mean of each of the multi-index attribute measure vectors. The risk quantification value is calculated based on the target attribute measurement vector and the maximum value of the endpoints of each measurement interval, and the card-freezing risk level of each card-freezing event type is determined based on the risk quantification value.

4. The method for early warning of machine jamming risk of tunnel boring machine according to claim 2, characterized in that, The process of determining the optimal and worst-case indicators for each risk indicator, and calculating the target subjective weight for each risk indicator based on the optimal and worst-case indicators, includes: Determine the set of evaluation indicators for each of the aforementioned risk indicators, and determine the best and worst indicators based on the set of evaluation indicators. Calculate the importance ratio of the optimal indicator and each evaluation indicator in the set of evaluation indicators to obtain the first comparison vector; Calculate the importance ratio of the worst-case indicator and each of the evaluation indicators in the set of evaluation indicators to obtain a second comparison vector; A subjective weight optimization function is constructed based on the best indicator, the worst indicator, the first comparison vector, and the second comparison vector. Set the boundary conditions for the subjective weight optimization function, calculate the subjective weight when the subjective weight optimization function reaches its minimum value, and obtain the target subjective weight.

5. The method for early warning of machine jamming risk of tunnel boring machine according to claim 2, characterized in that, Calculate the entropy value of each risk indicator based on the risk indicator data in historical card-related events, and determine the target objective weight of each risk indicator based on the entropy value, including: Obtain risk indicator data for each of the aforementioned risk indicators under multiple historical card-operated events, and construct an indicator data matrix based on each of the aforementioned risk indicators and the indicator data; The index data matrix is ​​standardized to obtain a standardized data matrix; Normalize each element of the standardized data matrix to obtain a standard normalized data matrix; The index entropy value of each risk indicator is calculated based on the standard normalized data matrix, and the objective weight is calculated based on the index entropy value to obtain the target objective weight.

6. The machine jam risk early warning method of a tunnel boring machine according to claim 2, characterized in that, The step of dynamically adjusting the combined weights based on the contribution scores of each risk indicator to the card machine risk, to obtain the dynamic risk assessment weights of each risk indicator, includes: Construct an indicator state vector based on the contribution score of each of the aforementioned risk indicators, and determine the balance function of the indicator state vector; The dynamic risk assessment weights are obtained by calculating the combined weights of each risk indicator and the normalized product of the indicator state vectors.

7. The method for early warning of machine jamming risk of tunnel boring machine according to claim 1, characterized in that, The step of calculating the error rate based on the cutter head feature data and cutter head threshold data, and determining the risk probability level of each of the jamming event types based on the error rate, includes: A tool turret threshold prediction model is constructed, and the tool turret threshold data is predicted based on the tool turret feature data and the target feature data using the tool turret threshold prediction model. The error rate is calculated based on the cutter head feature data and the cutter head threshold data, and the risk probability level is determined based on the error rate and the preset error threshold.

8. A jamming risk early warning system of a tunnel boring machine, characterized in that, The system includes: The machine jamming event classification module is used to classify the types of machine jamming events of tunnel boring machines based on historical event data, and to determine multiple machine jamming event types and corresponding preset machine jamming risk levels; The risk indicator determination module is used to identify geological risk sources in the target construction section of the tunnel boring machine and determine the risk indicators and quantitative standards for each type of machine jamming event. The data acquisition and processing module is used to acquire multi-source heterogeneous data of the tunnel boring machine, perform data preprocessing on the multi-source heterogeneous data, and obtain cutterhead feature data and target feature data corresponding to the risk indicators. The first risk prediction module is used to calculate the error rate based on the cutter head feature data and the target cutter head threshold, and to determine the risk probability level of each of the jamming event types based on the error rate. The second risk prediction module is used to calculate the dynamic risk assessment weight of each risk indicator in each of the card machine event types, and determine the card machine risk level of each of the card machine event types based on the dynamic risk assessment weight and the target feature data. The risk warning determination module is used to jointly determine the risk warning level of each of the card machine event types based on the risk probability level and the card machine risk level.

9. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, The processor executes the computer program to implement the tunnel boring machine jam risk warning method as described in any one of claims 1-7.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the tunnel boring machine jam risk warning method as described in any one of claims 1-7.

Images