Double-shield TBM split starting risk assessment method and system and storage medium

Abstract

The invention discloses a double-shield TBM split starting risk assessment method and system and a storage medium. The risk assessment method comprises the following steps that S1, accurate assembly and starting positioning of a TBM are achieved through a high-precision positioning module; s2, constructing a quantitative evaluation model of a double-shield TBM crossing key risk source, and carrying out quantitative evaluation on risk factors in a tunneling process; s3, performing fine quantitative analysis on the tunneling response through a tunneling response fine quantitative analysis module; s4, optimizing a tunneling cutter configuration and replacement strategy according to a tunneling response analysis result; and S5, establishing construction risk control and response measures, and ensuring the safety and quality stability of the construction process. Intelligent management of the double-shield TBM in the aspects of split starting, tunneling stable control, tunneling response quantitative analysis, construction risk control and the like is achieved, and it is ensured that tunneling engineering is completed safely, efficiently and with high quality.

Description

Technical Field

[0001] The present invention relates to the technical field of double-shield TBM, and in particular to a double-shield TBM split-launch risk assessment method, system and storage medium. Background Art

[0002] With the acceleration of global urbanization and the vigorous development of infrastructure construction, tunnel engineering, as an important engineering form to solve urban traffic congestion and cross complex terrain, has shown a significant growth trend in construction scale and technical difficulty. Especially in areas with complex and changeable geological conditions, high environmental sensitivity and high construction risks, such as crossing national railway lines, rivers, mountainous areas, etc., the performance and technical level of tunnel boring machines are directly related to the safety, quality and progress of the project.

[0003] As an important equipment in modern tunnel construction, double-shield full-face tunnel boring machines (TBMs) have been widely used in the construction of long and large tunnels due to their advantages such as fast excavation speed, good tunnel quality, and low environmental impact. However, in the face of key technical challenges such as split starting in a narrow space, excavation stability control under complex geological conditions, tool wear monitoring and replacement strategy optimization during excavation, traditional manual control and experience judgment can no longer meet the needs of modern tunnel construction.

[0004] The split start of a TBM in a narrow space requires extremely high assembly accuracy and operating efficiency to ensure the smooth start of the tunnel boring machine and the continuity of subsequent tunneling operations. At the same time, complex geological conditions such as faults, broken zones, and water gushing pose a huge challenge to the tunneling stability and safety of the tunnel boring machine. In addition, the wear of the cutter during the tunneling process directly affects the tunneling efficiency and tunneling cost, so it is necessary to monitor the wear status of the cutter in real time and formulate corresponding replacement strategies.

[0005] In order to solve the above problems, in recent years, with the rapid development of technologies such as intelligent perception, big data analysis, and artificial intelligence, the intelligence and information level of tunnel boring machines has been significantly improved. By integrating high-precision positioning, intelligent identification, real-time data analysis and other technical means, comprehensive monitoring and intelligent control of the excavation process can be achieved, improving excavation efficiency and safety. However, the intelligent application of double-shield TBM in terms of split start, excavation stability control, excavation response quantitative analysis, and construction risk control is still in its infancy, and further research and exploration is urgently needed. Summary of the invention

[0006] Based on the technical problems existing in the background technology, the present invention proposes a double-shield TBM split-starting risk assessment method, system and storage medium, which realizes the intelligent management of the double-shield TBM in split-starting, tunneling stability control, tunneling response quantitative analysis and construction risk control, ensuring the safe, efficient and high-quality completion of the tunnel excavation project.

[0007] The double shield TBM split launch risk assessment method proposed in the present invention has the following steps: S1: The high-precision positioning module is used to achieve precise assembly and launch positioning of the TBM; S2: Construct a quantitative assessment model for double-shield TBM crossing key risk sources and conduct a quantitative assessment of risk factors during tunneling; S3: Perform fine quantitative analysis on the excavation response through the excavation response fine quantitative analysis module; S4: Optimize tunneling tool configuration and replacement strategy based on tunneling response analysis results; S5: Establish construction risk control and response measures to ensure the safety and quality stability of the construction process.

[0008] Preferably, the method steps for the high-precision positioning module in S1 to perform positioning are as follows: S11: Obtain large-scale location information through GNSS global navigation, obtain short-term precise navigation information through INS inertial navigation, and obtain tunnel surrounding environment information through laser scanning; S12: Process the large-scale position information obtained by GNSS global navigation and the short-term precise navigation information obtained by INS inertial navigation through the Kalman filter to obtain the specific position data of the tunnel; S13: Obtaining positioning data based on the tunnel surrounding environment information and the tunnel specific location data; S14: Determine whether the positioning data matches the actual working condition. If so, output the positioning data. If not, adjust the initial data according to the actual working condition.

[0009] Preferably, the method steps for quantitatively evaluating risk factors in the excavation process in S2 are as follows: S21: Identify and determine key risk sources based on geological exploration data, historical construction cases and on-site expert experience; S22: Classify key risk sources; S23: obtaining geological parameters, excavation parameters and construction environment parameters through sensors; S24: Establish a quantitative assessment model based on risk sources and obtained parameters; S25: Analyze real-time monitoring data and historical monitoring data through quantitative assessment models and output risk levels and potential impact ranges.

[0010] Preferably, the method steps for establishing the quantitative evaluation model in S24 are as follows: S241: The tunnel engineering calculation model is built by integrating the foundation parameters and support material parameters of the construction section under different conditions through numerical analysis software; S242: training a tunnel engineering calculation model according to parameters involved in the tunnel engineering to determine a prediction model; S243: Determine information gain based on parameters involved in tunnel engineering: In the formula, is the information gain ratio; is the information gain; is the ratio parameter; l is the partition variable; L is the total number of variables; D is the total number of parameters in the training set; D l is the parameter component in the training set; p k is the sample proportion; is information entropy; k is the sample individual; y is the total number of samples.

[0011] S244: The prediction model in S242 constructs a decision tree model based on information gain, trains the decision tree model through a training data set, and determines the splitting rule of each feature node; S245: Evaluate the trained model to complete the model establishment.

[0012] Preferably, the accuracy calculation formula of the model evaluation in S245 is: In the formula, ACC is the accuracy of the evaluation model; is the sample individual set; is a binary indicator function; is the true label of the sample; is the predicted label of the sample.

[0013] Preferably, the tunneling response fine quantitative analysis module in S3 includes: The data acquisition module uploads the excavation parameter APP log data in real time through TBM sensors; it is used for daily excavation incremental data transmission and equipment log parameter transmission; it is used for DataX excavation parameter synchronization; The data processing module sequentially performs excavation data cleaning and integration, excavation data preprocessing, excavation parameter feature extraction and pattern recognition, and nonlinear data feature relationship analysis; The analysis module analyzes the results of the data processing module through the quantitative evaluation model and optimizes the quantitative evaluation model; Decision support module for data management and maintenance.

[0014] Preferably, the method steps for optimizing the tunneling tool configuration and replacement strategy in S4 are as follows: S41: Acquire tool status data through the TBM tool sensor and transmit it to the tool status data analysis module; S42: The tool status data analysis module evaluates the tool performance and determines the optimization plan for the tool configuration.

[0015] The double-shield TBM split launching risk assessment system proposed by the present invention comprises: Precision positioning module, used to achieve precise assembly and launch positioning of TBM; The risk assessment module is used to build a quantitative assessment model for double-shield TBM crossing key risk sources and conduct a quantitative assessment of risk factors during the tunneling process; The module for fine quantitative analysis of tunneling response is used to conduct fine quantitative analysis of tunneling response; Tool status monitoring module, used to optimize tunneling tool configuration and replacement strategy based on tunneling response analysis results; The risk control module is used to establish construction risk control and response measures to ensure the safety and quality stability of the construction process.

[0016] The computer-readable storage medium proposed in the present invention stores a computer program, which is characterized in that when the computer program is executed by a processor, the above-mentioned double-shield TBM split-initiation risk assessment method is implemented.

[0017] Beneficial technical effects of the present invention: The present invention realizes real-time high-precision positioning of the TBM tunneling machine through a high-precision positioning module. By integrating data from multiple positioning sources, it effectively reduces error accumulation, improves positioning accuracy and stability, and provides a reliable basis for precise control of the tunneling process.

[0018] The present invention realizes risk control and prediction of potential impact range of TBM initiation and excavation process through risk assessment module. The prediction model given by machine learning algorithm effectively improves the safety of the project, reduces construction risks, and provides a reliable basis for excavation construction decision-making. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] Figure 1 A flowchart of positioning by the high-precision positioning module proposed by the present invention; Figure 2 A flow chart for quantitatively evaluating risk factors in the excavation process proposed by the present invention; Figure 3 It is a schematic diagram of the fine quantitative analysis module of tunneling response proposed by the present invention; Figure 4 A flow chart of the optimization of tunneling tool configuration and replacement strategy proposed by the present invention; Figure 5 This is a flow chart for establishing construction risk control and response measures proposed by the present invention. DETAILED DESCRIPTION

[0020] The present invention will be further explained below in conjunction with specific embodiments.

[0021] Example 1 The double shield TBM split launch risk assessment method proposed in the present invention has the following steps: S1: The high-precision positioning module is used to achieve precise assembly and launch positioning of the TBM; S2: Construct a quantitative assessment model for double-shield TBM crossing key risk sources and conduct a quantitative assessment of risk factors during tunneling; S3: Perform fine quantitative analysis on the excavation response through the excavation response fine quantitative analysis module; S4: Optimize tunneling tool configuration and replacement strategy based on tunneling response analysis results; S5: Establish construction risk control and response measures to ensure the safety and quality stability of the construction process.

[0022] Specifically, refer to Figure 1 The steps for positioning by the high-precision positioning module in S1 are as follows: S11: Obtain large-scale location information through GNSS global navigation, obtain short-term precise navigation information through INS inertial navigation, and obtain tunnel surrounding environment information through laser scanning; S12: Process the large-scale position information obtained by GNSS global navigation and the short-term precise navigation information obtained by INS inertial navigation through the Kalman filter to obtain the specific position data of the tunnel; S13: Obtaining positioning data based on the tunnel surrounding environment information and the tunnel specific location data; S14: Determine whether the positioning data matches the actual working condition. If so, output the positioning data. If not, adjust the initial data according to the actual working condition.

[0023] As a preferred embodiment of the present invention, refer to Figure 2 , the method steps for quantitatively evaluating the risk factors in the tunneling process in S2 are as follows: S21: Identify and determine key risk sources based on geological exploration data, historical construction cases and on-site expert experience; S22: Classify key risk sources, including national railway lines, fault zones and weak strata; S23: obtaining geological parameters, excavation parameters and construction environment parameters through sensors; the geological parameters mainly include lithology data, formation dip data and fault distribution data; the excavation parameters mainly include excavation speed parameters, cutterhead speed parameters and thrust parameters; the construction environment parameters mainly include groundwater level parameters, temperature change parameters and other geological environment parameters; S24: Establish a quantitative assessment model based on risk sources and obtained parameters; S25: Analyze real-time monitoring data and historical monitoring data through quantitative assessment models and output risk levels and potential impact ranges.

[0024] The method steps for establishing the quantitative evaluation model in S24 are as follows: S241: The tunnel engineering calculation model is built by integrating the foundation parameters and support material parameters of the construction sections (ordinary sections and risk sections) under different conditions through numerical analysis software such as FLAC3D and MIDASGTSNX; S242: Relying on the numerical analysis and processing computing capabilities of the analysis software, key parameters such as cutter head speed, thrust, horizontal displacement, longitudinal displacement, and crown bending moment are collected; these parameters will be input into the machine learning model in groups, and by transmitting engineering data to different types of learning models such as decision trees, random forests, artificial neural networks, and Bayesian learning for a period of time, tunnel engineering calculation models trained under different algorithm modes can be obtained. These models can evaluate the safety performance of tunnel TBM construction under different conditions (such as different TBM cutter head speeds and thrusts, different geological soil conditions or special risk formations), and give the scope and degree of influence of tunnel excavation on surrounding formations; after professional engineers select these tunnel engineering calculation models, we can finally obtain the prediction model with the highest degree of fit with the current tunnel engineering. In this way, we only need to input various monitoring data, exploration data, and historical data of the current tunnel engineering into the best prediction model, and we can obtain construction risk level data through the analysis of the prediction model, which provides a reference for subsequent construction decisions; S243: Determine the information gain based on the parameters involved in the tunnel project, that is, the degree to which the uncertainty of the data set is reduced by the influencing factor: In the formula, is the information gain ratio; is the information gain; is the ratio parameter; l is the partition variable; L is the total number of variables; D is the total number of parameters in the training set; D l is the parameter component in the training set; p k is the sample proportion; is information entropy; k is the sample individual; y is the total number of samples.

[0025] S244: The prediction model in S242 constructs a decision tree model based on information gain, trains the decision tree model through a training data set, and determines the splitting rules of each feature node. For example, the tunnel crown bending moment value is split into two nodes based on whether it exceeds the limit value, leading to a high-scoring path and a low-scoring path respectively; S245: Then use the DecisionTreeClassifier in the scikit-learn library to train the collected data, and optimize the model through boundary conditions such as max_depth and min_samples_split to prevent overfitting; for the trained model, use the test data set to evaluate the model, mainly comparing the safety level corresponding to the score output by the model with the corresponding score range of the safety level under actual engineering conditions, and calculate the model's accuracy, recall rate, F1 score and other indicators to complete the model establishment.

[0026] The accuracy calculation formula of the model evaluation in S245 is: In the formula, ACC is the accuracy of the evaluation model; is the sample individual set; is a binary indicator function; is the true label of the sample; is the predicted label of the sample.

[0027] After the model is established, it is applied to new construction data to predict the risks that may arise during the construction process, and corresponding risk control measures are formulated based on the risk level output by the model. Finally, during the construction process, new monitoring data is collected in real time, such as microseismic monitoring data, surface settlement data, etc., and the model is dynamically updated to reflect changes in the construction process and improve the accuracy of risk prediction.

[0028] Reference Figure 3 The fine quantitative analysis module of tunneling response in S3 includes: The data acquisition module uploads the excavation parameter APP log data in real time through TBM sensors; it is used for daily excavation incremental data transmission and equipment log parameter transmission; it is used for DataX excavation parameter synchronization; The data processing module sequentially performs excavation data cleaning and integration, excavation data preprocessing, excavation parameter feature extraction and pattern recognition, and nonlinear data feature relationship analysis; The analysis module analyzes the results of the data processing module through the quantitative evaluation model and optimizes the quantitative evaluation model. The optimization of the quantitative evaluation model mainly optimizes the data analysis results through BP neural network, SVW support vector machine or RF random forest, and performs goodness of fit evaluation and MAPE indicator evaluation on the optimized model. The decision support module is used for data management and maintenance, and mainly includes data management subsystem, OKPS subsystem, auxiliary function subsystem, decision subsystem and system maintenance subsystem.

[0029] Reference Figure 4 The method steps for optimizing the tunneling tool configuration and replacement strategy in S4 are as follows: S41: Obtain tool status data through the TBM tool sensor and transmit it to the tool status data analysis module; the tool status display interface mainly includes the degree of cutter head wear, cutter head speed, torque and thrust, cutter head vibration acceleration analysis, cutter head FFT vibration frequency analysis, cutter head penetration depth and cutter head working environment parameters; S42: The tool status data analysis module evaluates the tool performance and determines the optimization plan for tool configuration; the performance evaluation mainly adopts engineering experience analogy evaluation and expert database evaluation; the tool configuration optimization plan mainly includes tool type configuration plan, tool quantity optimization plan, worn tool replacement plan and tool performance endurance plan.

[0030] Reference Figure 5 The establishment of construction risk control and response measures in S5 mainly includes: risk identification and assessment, risk classification management, formulation of risk prevention measures, formulation of risk monitoring and early warning system, formulation of emergency response plan, risk control and decision support system, risk overall coordination system, risk follow-up management system, construction environment maintenance system, quality supervision and safety monitoring system.

[0031] Example 2 The double-shield TBM split launching risk assessment system proposed by the present invention comprises: Precision positioning module, used to achieve precise assembly and launch positioning of TBM; The risk assessment module is used to build a quantitative assessment model for double-shield TBM crossing key risk sources and conduct a quantitative assessment of risk factors during the tunneling process; The module for fine quantitative analysis of tunneling response is used to conduct fine quantitative analysis of tunneling response; Tool status monitoring module, used to optimize tunneling tool configuration and replacement strategy based on tunneling response analysis results; The risk control module is used to establish construction risk control and response measures to ensure the safety and quality stability of the construction process.

[0032] Example 3 The computer-readable storage medium proposed in the present invention stores a computer program, which is characterized in that when the computer program is executed by a processor, the double-shield TBM split-initiation risk assessment method in Example 1 is implemented.

Claims

1. Double shield TBM split initial risk assessment method, characterized by: The steps are as follows: S1: The high-precision positioning module is used to achieve precise assembly and launch positioning of the TBM; S2: Construct a quantitative assessment model for double-shield TBM crossing key risk sources and conduct a quantitative assessment of risk factors during tunneling; S3: Perform fine quantitative analysis on the excavation response through the excavation response fine quantitative analysis module; S4: Optimize tunneling tool configuration and replacement strategy based on tunneling response analysis results; S5: Establish construction risk control and response measures to ensure the safety and quality stability of the construction process.

2. The double shield TBM split launch risk assessment method according to claim 1 is characterized in that: The steps for positioning by the high-precision positioning module in S1 are as follows: S11: Obtain large-scale location information through GNSS global navigation, obtain short-term precise navigation information through INS inertial navigation, and obtain tunnel surrounding environment information through laser scanning; S12: Process the large-scale position information obtained by GNSS global navigation and the short-term precise navigation information obtained by INS inertial navigation through the Kalman filter to obtain the specific position data of the tunnel; S13: Obtaining positioning data based on the tunnel surrounding environment information and the tunnel specific location data; S14: Determine whether the positioning data matches the actual working condition. If so, output the positioning data. If not, adjust the initial data according to the actual working condition.

3. The double shield TBM split launch risk assessment method according to claim 1 is characterized in that: The method steps for quantitatively evaluating risk factors during excavation in S2 are as follows: S21: Identify and determine key risk sources based on geological exploration data, historical construction cases and on-site expert experience; S22: Classify key risk sources; S23: obtaining geological parameters, excavation parameters and construction environment parameters through sensors; S24: Establish a quantitative assessment model based on risk sources and obtained parameters; S25: Analyze real-time monitoring data and historical monitoring data through quantitative assessment models and output risk levels and potential impact ranges.

4. The double shield TBM split launch risk assessment method according to claim 3 is characterized in that: The method steps for establishing a quantitative evaluation model in S24 are as follows: S241: The tunnel engineering calculation model is built by integrating the foundation parameters and support material parameters of the construction section under different conditions through numerical analysis software; S242: training a tunnel engineering calculation model according to parameters involved in the tunnel engineering to determine a prediction model; S243: Determine information gain based on parameters involved in tunnel engineering: Where, Gain rating(D,l) is the information gain ratio; Gain(D,l) is the information gain; C(l) is the ratio parameter; l is the partition variable; L is the total number of variables; D is the total number of parameters in the training set; D l is the parameter component in the training set; p k is the sample proportion; E(D) is the information entropy; k is the sample individual; y is the total number of samples; S244: The prediction model in S242 constructs a decision tree model based on information gain, trains the decision tree model through a training data set, and determines the splitting rule of each feature node; S245: Evaluate the trained model to complete the model establishment.

5. The double shield TBM split launch risk assessment method according to claim 4 is characterized in that: The accuracy calculation formula of the model evaluation in S245 is: Where ACC is the accuracy of the evaluation model; (x, y) is the sample individual set; T(y) is the binary indicator function; y is the true label of the sample; is the predicted label of the sample.

6. The double shield TBM split launch risk assessment method according to claim 1, characterized in that: The fine quantitative analysis module of tunneling response in S3 includes: The data acquisition module uploads the excavation parameter APP log data in real time through TBM sensors; it is used for daily excavation incremental data transmission and equipment log parameter transmission; it is used for DataX excavation parameter synchronization; The data processing module sequentially performs excavation data cleaning and integration, excavation data preprocessing, excavation parameter feature extraction and pattern recognition, and nonlinear data feature relationship analysis; The analysis module analyzes the results of the data processing module through the quantitative evaluation model and optimizes the quantitative evaluation model; Decision support module for data management and maintenance.

7. The double shield TBM split launch risk assessment method according to claim 1, characterized in that: The method steps for optimizing the tunneling tool configuration and replacement strategy in S4 are as follows: S41: Acquire tool status data through the TBM tool sensor and transmit it to the tool status data analysis module; S42: The tool status data analysis module evaluates the tool performance and determines the optimization plan for the tool configuration.

8. Double shield TBM split launch risk assessment system, characterized by: include: Precision positioning module, used to achieve precise assembly and launch positioning of TBM; The risk assessment module is used to build a quantitative assessment model for double-shield TBM crossing key risk sources and conduct a quantitative assessment of risk factors during the tunneling process; The module for fine quantitative analysis of tunneling response is used to conduct fine quantitative analysis of tunneling response; Tool status monitoring module, used to optimize tunneling tool configuration and replacement strategy based on tunneling response analysis results; The risk control module is used to establish construction risk control and response measures to ensure the safety and quality stability of the construction process.

9. A computer-readable storage medium storing a computer program, characterized in that: When the computer program is executed by a processor, the double-shield TBM split-initiation risk assessment method as described in any one of claims 1 to 7 is implemented.

Images