TBM tunnel face surrounding rock grade real-time intelligent sensing method and system

Abstract

The application provides a TBM tunnel face surrounding rock grade real-time intelligent sensing method and system, the method comprises the following steps: collecting rock-machine related multi-source information in the TBM tunneling process; establishing a TBM tunnel face surrounding rock classification standard; based on the TBM tunnel face surrounding rock classification standard, a database for training a rock-machine composite score cascade prediction model is constructed; taking the constructed database as a training sample, a rock-machine composite score cascade prediction model is constructed based on a ridge regression model and an artificial neural network; inputting the rock-machine related multi-source information into the rock-machine composite score cascade prediction model, obtaining the current TBM tunnel face composite rock-machine score, and then obtaining the TBM tunnel face surrounding rock grade according to the TBM tunnel face surrounding rock classification standard. The application realizes real-time intelligent prediction of the TBM tunnel face surrounding rock grade, improves the TBM tunneling environment perception level, and lays a solid foundation for underground engineering intelligent construction.

Description

Technical Field

[0001] This application relates to the field of underground engineering construction technology, specifically to a real-time intelligent sensing method and system for the surrounding rock grade of a TBM tunnel face during excavation. Background Technology

[0002] TBM (Tunnel Boring Machine) tunnel construction is a highly mechanized tunnel construction method. Compared with drill and blast, TBM has many advantages such as less disturbance to the surrounding rock and higher excavation efficiency, and has been widely used in tunnel engineering construction. How to quickly and accurately determine the surrounding rock grade at the TBM tunnel face is of great significance for optimizing TBM tunneling performance.

[0003] Currently, predictions are mainly made through advanced geological forecasting or TBM operation parameter analysis. However, advanced geological forecasting is complex and time-consuming, and TBM operation parameters are greatly affected by TBM driving behavior and have obvious randomness and dispersion. In addition, the existing surrounding rock classification standards are not specifically designed for TBM tunnels. They only consider the single factor of surrounding rock geological conditions and ignore the TBM's own tunneling performance. Summary of the Invention

[0004] The TBM tunnel face surrounding rock grade real-time intelligent sensing method provided in this application can solve the problems of complex and time-consuming advanced geological prediction operation in the existing technology for sensing the surrounding rock grade of TBM tunnel face, and the fact that it only considers the single factor of surrounding rock geological conditions and ignores the TBM's own tunneling performance.

[0005] Firstly, this application provides a method for real-time intelligent sensing of the surrounding rock grade at the tunnel face during TBM excavation, comprising the following steps:

[0006] Collect and acquire multi-source information related to the rock machine during the TBM tunneling process. The multi-source information related to the rock machine includes multi-source information related to the TBM and the surrounding rock.

[0007] Establish a classification standard for the surrounding rock at the working face of a TBM tunnel;

[0008] Based on the TBM tunnel face surrounding rock classification standard, a database is constructed for training the rock machine composite score cascade prediction model;

[0009] Using the constructed database as training samples, a rock machine composite score cascade prediction model was constructed based on the ridge regression model and artificial neural network.

[0010] Input the multi-source information related to rock machinery during the TBM tunneling process into the constructed rock machinery composite score cascade prediction model to obtain the composite rock machinery score of the current TBM tunnel face. Then, based on the constructed TBM tunnel face surrounding rock classification standard, obtain the TBM tunnel face surrounding rock grade.

[0011] In conjunction with the first aspect, in one implementation method, the acquisition of multi-source information related to the rock machine during the TBM tunneling process specifically includes the following steps:

[0012] Collect information on the interaction between the TBM and the rock excavator during the tunneling process;

[0013] Collect information on the morphology of rock debris during TBM tunneling;

[0014] Collect current conduction information during the TBM tunneling process;

[0015] Collect information on rock-breaking dust during the TBM tunneling process.

[0016] In conjunction with the first aspect, in one implementation method, the establishment of a grading standard for the surrounding rock at the TBM tunnel face specifically includes the following steps:

[0017] Determine the score for the surrounding rock;

[0018] Determine the tunneling score;

[0019] The weighting coefficients of the surrounding rock score and the tunneling score are determined by the analytic hierarchy process (AHP), and the weighted sum of the surrounding rock score and the tunneling score is used to calculate the rock-machine composite score.

[0020] Based on the rock-mechanical composite score, a classification standard for the surrounding rock at the tunnel face of a TBM is established.

[0021] In conjunction with the first aspect, in one implementation, the step of constructing a database for training a rock-machine composite score cascade prediction model based on the TBM tunnel face surrounding rock grading standard specifically includes the following steps:

[0022] Collect multi-source information related to rock excavation machines during the TBM tunneling process from excavated sections of multiple TBM projects;

[0023] The collected multi-source information related to rock machines during the TBM tunneling process is compared with the constructed TBM tunnel face surrounding rock classification standard to obtain a composite rock machine score.

[0024] Using the collected multi-source information related to rock machines during the TBM tunneling process as input and the corresponding composite rock machine score as target output, a database including the input and target output is constructed for training a cascaded prediction model of rock machine scores.

[0025] In conjunction with the first aspect, in one implementation, the step of using the constructed database as training samples to construct a rock machine composite score cascade prediction model based on a ridge regression model and an artificial neural network specifically includes the following steps:

[0026] Multiple training sets are obtained by sampling the database with replacement, wherein the size of the training sets is the same as the size of the database.

[0027] Ridge regression models were constructed using multiple training sets to serve as the first-level rock machine composite score prediction model.

[0028] Input the training samples from the database into the first-level rock machine composite score prediction model to obtain the first-level prediction results;

[0029] The first-level prediction results are combined and used as the input of the updated training samples while keeping the target output unchanged. An artificial neural network model is constructed based on the updated database to build a second-level rock machine composite score prediction model, and the output of the second-level rock machine composite score prediction model is used as the final prediction result.

[0030] In conjunction with the first aspect, in one implementation, in the step of constructing ridge regression models using multiple training sets as the first-level rock machine composite score prediction model, the objective function f(w) of the ridge regression model is as follows:

[0031] f(w) = ||Xw - y||2 2 +α||w||2 2

[0032] In the formula, X refers to the input vector of the training sample; y refers to the target output of the training sample; ||·||2 refers to the Euclidean norm; and w refers to the coefficients to be solved in the ridge regression model.

[0033] In conjunction with the first aspect, in one implementation, the loss function Loss of the artificial neural network model is as follows:

[0034]

[0035] In the formula, weight and bias are the neuron connection weights and biases in the artificial neural network model, respectively.

[0036] In conjunction with the first aspect, in one implementation, the following steps are also included:

[0037] Get the test set;

[0038] Based on the test set, the perception performance of the rock machine composite score prediction model is calculated and the model's perception performance is verified.

[0039] Secondly, this application provides a real-time intelligent sensing system for the surrounding rock grade at the tunnel face during TBM excavation, comprising:

[0040] The multi-source information acquisition module is used to collect and acquire multi-source information related to the rock machine during the TBM tunneling process;

[0041] The grading standard construction module is used to establish grading standards for the surrounding rock of TBM tunnel faces;

[0042] The database construction module is communicatively connected to the grading standard construction module and is used to construct a database for training the rock machine composite score cascade prediction model based on the TBM tunnel face surrounding rock grading standard.

[0043] The prediction model building module communicates with the database building module and is used to build a rock machine composite score cascade prediction model based on the ridge regression model and artificial neural network, using the built database as training samples.

[0044] The tunnel face surrounding rock grade acquisition module communicates with the multi-source information acquisition module, the prediction model construction module, and the grading standard construction module. It is used to input the multi-source information related to rock machinery during the TBM tunneling process into the constructed rock machinery composite score cascade prediction model to obtain the composite rock machinery score of the current TBM tunnel face. Then, based on the constructed TBM tunnel face surrounding rock grading standard, the TBM tunnel face surrounding rock grade is obtained.

[0045] In conjunction with the second aspect, in one implementation, the multi-source information acquisition module includes:

[0046] The rock-machine interaction information acquisition unit is used to collect rock-machine interaction information during the TBM tunneling process;

[0047] Rock debris morphology information acquisition unit, used to collect rock debris morphology information during TBM tunneling;

[0048] The current conduction information acquisition unit is used to collect current conduction information during the TBM tunneling process;

[0049] The rock-breaking dust information acquisition unit is used to collect rock-breaking dust information during the TBM tunneling process.

[0050] The beneficial effects of the technical solutions provided in this application include at least the following:

[0051] This application provides a real-time intelligent perception method for the surrounding rock grade at the tunnel face of a TBM tunnel. By collecting multi-source information related to the rock excavator during excavation and constructing a joint prediction model for the surrounding rock grade suitable for TBM tunnels based on ridge regression and artificial neural networks, the method achieves real-time intelligent prediction of the surrounding rock grade at the tunnel face of a TBM tunnel, improves the TBM's perception level of the tunneling environment, and lays a solid foundation for intelligent construction of underground engineering. Attached Figure Description

[0052] Figure 1 A flowchart illustrating the real-time intelligent sensing method for the surrounding rock grade of a TBM tunnel face provided in this embodiment of the application;

[0053] Figure 2 This is a schematic diagram of the rock-machine interaction information acquisition process provided in the embodiments of this application;

[0054] Figure 3 This is a schematic flowchart of a method for collecting rock slag morphology information provided in an embodiment of this application;

[0055] Figure 4 This is a schematic flowchart of a method for collecting current conduction information provided in an embodiment of this application;

[0056] Figure 5 This is a schematic flowchart of a method for collecting rock-breaking dust information provided in an embodiment of this application;

[0057] Figure 6 This is a schematic diagram of the process for constructing a TBM tunnel surrounding rock classification standard provided in an embodiment of this application;

[0058] Figure 7 This is a schematic diagram of a method for constructing a database provided in an embodiment of this application;

[0059] Figure 8 This is a schematic diagram of the method for constructing a rock machine composite score cascade prediction model provided in an embodiment of this application;

[0060] Figure 9 A schematic flowchart of a method for sensing the surrounding rock grade of a TBM face during excavation, provided in an embodiment of this application;

[0061] Figure 10 This is a schematic diagram illustrating the comparative analysis of the sensing performance provided in an embodiment of this application. Detailed Implementation

[0062] To enable those skilled in the art to better understand the present application, the technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present application, and not all embodiments. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of the present application.

[0063] The terms "comprising" and "having," and any variations thereof, in the specification, claims, and accompanying drawings of this application are intended to cover non-exclusive inclusion. For example, a process, method, system, product, or apparatus that includes a series of steps or units is not limited to the listed steps or units, but may optionally include steps or units not listed, or may optionally include other steps or units inherent to such process, method, product, or apparatus. The terms "first," "second," and "third," etc., are used to distinguish different objects, etc., and do not indicate a sequence, nor do they limit "first," "second," and "third" to different types.

[0064] In the description of the embodiments of this application, terms such as "exemplary," "for example," or "for instance" are used to indicate examples, illustrations, or explanations. Any embodiment or design described as "exemplary," "for example," or "for instance" in the embodiments of this application should not be construed as being more preferred or advantageous than other embodiments or designs. Specifically, the use of terms such as "exemplary," "for example," or "for instance" is intended to present the relevant concepts in a concrete manner.

[0065] In the description of the embodiments of this application, unless otherwise stated, " / " means "or". For example, A / B can mean A or B. The "and / or" in the text is merely a description of the relationship between related objects, indicating that there can be three relationships. For example, A and / or B can mean: A exists alone, A and B exist simultaneously, and B exists alone. In addition, in the description of the embodiments of this application, "multiple" means two or more.

[0066] In some processes described in the embodiments of this application, multiple operations or steps are included in a specific order. However, it should be understood that these operations or steps may not be executed in the order they appear in the embodiments of this application, or they may be executed in parallel. The sequence number of the operation is only used to distinguish different operations, and the sequence number itself does not represent any execution order. In addition, these processes may include more or fewer operations, and these operations or steps may be executed sequentially or in parallel, and these operations or steps may be combined.

[0067] To make the objectives, technical solutions, and advantages of this application clearer, the embodiments of this application will be described in further detail below with reference to the accompanying drawings.

[0068] Firstly, please refer to Figure 1 This application provides a real-time intelligent sensing method for the surrounding rock grade at the tunnel face during TBM excavation, comprising the following steps:

[0069] Step S1: Collect and acquire multi-source information related to the rock machine during the TBM tunneling process. The multi-source information related to the rock machine includes multi-source information related to the TBM and the surrounding rock.

[0070] Step S2: Establish a classification standard for the surrounding rock at the TBM tunnel face;

[0071] Step S3: Based on the TBM tunnel face surrounding rock classification standard, construct a database for training the rock machine composite score cascade prediction model;

[0072] Step S4: Using the constructed database as training samples, construct a rock machine composite score cascade prediction model based on ridge regression model and artificial neural network;

[0073] Step S5: Input the multi-source information related to rock machinery during the TBM tunneling process into the constructed rock machinery composite score cascade prediction model to obtain the composite rock machinery score of the current TBM tunnel face. Then, based on the constructed TBM tunnel face surrounding rock classification standard, obtain the TBM tunnel face surrounding rock grade.

[0074] This application provides a real-time intelligent perception method for the surrounding rock grade at the tunnel face of a TBM tunnel. By collecting multi-source information related to the rock excavator during excavation and constructing a joint prediction model for the surrounding rock grade suitable for TBM tunnels based on ridge regression and artificial neural networks, the method achieves real-time intelligent prediction of the surrounding rock grade at the tunnel face of a TBM tunnel, improves the TBM's perception level of the tunneling environment, and lays a solid foundation for intelligent construction of underground engineering.

[0075] In one embodiment, step S1: acquiring multi-source information related to the rock excavator during the TBM tunneling process specifically includes the following steps:

[0076] Step S1A: Collect rock-machine interaction information during TBM tunneling; specifically, TBM operating parameters contain rich rock-machine interaction information. Through the TBM's onboard sensing system, operating parameters such as cutterhead thrust, cutterhead torque, and cutterhead penetration are monitored in real time during TBM tunneling. Based on the TBM operating parameters collected during tunneling monitoring, on-site penetration index, torque penetration index, and rock breaking ratio are extracted simultaneously as representative features of the rock-machine interaction information; more specifically, such as... Figure 2The TBM's operating parameters, such as cutterhead thrust, cutterhead torque, and cutterhead penetration, are collected during excavation using its own onboard sensing system. These parameters are then numerically transformed according to equations (1) to (3) to calculate representative features of the rock-machine interaction information: on-site penetration index, torque penetration index, and rock breaking ratio.

[0077]

[0078] In the formula, FPI, TPI, and SE refer to the field penetration index, torque penetration index, and rock breaking energy, respectively; F, T, and PR refer to the cutterhead thrust, cutterhead torque, and cutterhead penetration, respectively; N refers to the number of roller cutters on the cutterhead; and R refers to the cutterhead diameter.

[0079] Step S1B: Collect rock debris morphology information during TBM tunneling to reflect the TBM's rock breaking situation; specifically, in order to continuously collect rock debris morphology information in real time, such as... Figure 3 As shown, a rock debris morphology information acquisition device is installed. For example, a camera is placed directly above the rock debris conveyor belt connected to the slag outlet of the cutterhead screw conveyor. During excavation, the two-dimensional projection morphology parameters of the rock debris formed on the conveyor belt plane under the rock breaking action of the TBM are acquired, namely the contour perimeter and contour area. Then, the two-dimensional projection morphology parameters of the rock debris are synchronously mapped to the three-dimensional solid morphology parameters of the rock debris, namely the three-dimensional volume and three-dimensional sphericity of the rock debris, through geometric transformation.

[0080] More specifically, step S1B includes the following steps:

[0081] A camera is positioned directly above the rock slag conveyor belt connected to the slag outlet of the cutterhead screw conveyor, with the camera height 60cm above the conveyor belt plane.

[0082] The two-dimensional projection morphological parameters, i.e., the perimeter P, of the rock debris collected during excavation on the conveyor belt plane are as follows: 2D and contour area A 2D ;

[0083] The equivalent diameter D of the two-dimensional projection of the rock slag is calculated using equations (4) and (5), respectively. 2D and S sphere degree S 2D :

[0084]

[0085] The morphological parameters of the rock debris entity, i.e., the three-dimensional volume V, are calculated using equations (6) and (7). 3D and three-dimensional sphericity S 3D :

[0086] V 3D =3π(D) 2D ) 3 / 25 Formula (6)

[0087] S 3D =2.5S 2D -1.525 Equation (7)

[0088] Step S1C: As Figure 4 As shown, current conduction information during TBM tunneling is collected. The change in current under equal pressure reflects the change in resistance of the conduction medium, and the magnitude of resistance is closely related to the rock mass structure and rock properties of the surrounding rock. Specifically, in order to collect current conduction information of the surrounding rock at the tunnel face in real time and continuously, an electrode is installed in the cutter groove on both sides of each TBM cutter head, and a voltage stabilizing power supply device and a current detection device are placed on the back of the cutter head. Before the TBM contacts the surrounding rock, the current passing through the cutter head is monitored and used as a reference value. Then, the current during the TBM rock breaking process is monitored during tunneling, and the difference between the current and the aforementioned reference value is calculated simultaneously as the current conduction characteristic of the surrounding rock at the tunnel face.

[0089] More specifically, step S1C includes the following steps:

[0090] Step S1C1: Install an electrode in each of the two slots on both sides of each hob, and place the regulated power supply and ammeter on the back of the cutter head. Pass the electrode leads back to the back of the cutter head through the slot openings to form a complete current loop with the regulated power supply and ammeter.

[0091] Step S1C2: Measure the current I0 passing through the cutter before the TBM comes into contact with the surrounding rock and use it as a reference value.

[0092] Step S1C3: Monitor the current I during the TBM rock breaking process during excavation. t Simultaneously calculate the difference ΔI between it and the aforementioned benchmark value I0. t As a characteristic of current conduction in the surrounding rock of the tunnel face, the formula for calculating the difference is as follows:

[0093] ΔI t =I t -I0.

[0094] Step S1D: Collect rock-breaking dust information during TBM tunneling; Under the penetration and cutting action of the TBM cutterhead, the surrounding rock at the tunnel face will be damaged and dust will be generated. The rock-breaking dust information is evaluated and obtained based on the particle size of the dust generated during rock breaking at the tunnel face; specifically, such as... Figure 5 As shown, a rock-breaking dust information collection device is installed, such as a dust particle size distribution measuring instrument installed at the slag inlet of the cutterhead screw conveyor, to collect the dust particle size distribution generated by rock breaking at the working face, divide the dust particle size generated by rock breaking at the working face into different interval particle size ranges, and use the non-uniformity coefficient of dust quality in different interval particle size ranges as the characteristics of rock breaking dust at the working face.

[0095] More specifically, step S1D includes the following steps:

[0096] Step S1D1: Install a dust particle size distribution meter at the slag inlet of the screw conveyor and measure the mass of dust particles with particle sizes >150μm, 100~150μm, 50~100μm and <50μm respectively.

[0097] Step S1D2: Monitoring the mass of dust of different particle sizes generated during the TBM rock breaking process during excavation. and Simultaneously calculate the difference between its monitoring result and the previous monitoring result. The mass difference of dust with different particle sizes This information represents the dust generated by rock breaking at the current TBM tunnel face, where i is replaced by >150, 100~150, 50~100, and <50, respectively. Specifically, wait.

[0098] Step S1D3: Based on the mass difference of dust particles of different sizes, calculate the non-uniformity coefficient UC of dust particles of different sizes using equation (8), and use it as the characteristic of rock-breaking dust in the working face:

[0099]

[0100] In the formula, Mean t refer to and The average value.

[0101] In one embodiment, such as Figure 6 As shown, step S2: establishing a grading standard for the surrounding rock at the TBM tunnel face, specifically includes the following steps:

[0102] Step S21: Determine the surrounding rock score; specifically: the surrounding rock score is determined by the international rock grade, which is divided into Grade I, Grade II, Grade IV and Grade III surrounding rock; more specifically: Grade III surrounding rock is the most favorable for TBM tunneling, and the corresponding surrounding rock score is recorded as 5; Grade II and Grade IV surrounding rock are next, and the corresponding surrounding rock score is recorded as 3; Grade I and Grade V surrounding rock are the worst, and the corresponding surrounding rock score for Grade I and Grade V surrounding rock is recorded as 1.

[0103] Step S22: Determine the tunneling score; specifically, the tunneling score is determined by the TBM's net tunneling rate; more specifically: a TBM net tunneling rate > 80 mm / min corresponds to a tunneling score of 5; a TBM net tunneling rate between 60 and 80 mm / min corresponds to a tunneling score of 4; a TBM net tunneling rate between 40 and 60 mm / min corresponds to a tunneling score of 3; a TBM net tunneling rate between 20 and 40 mm / min corresponds to a tunneling score of 2; a TBM net tunneling rate < 20 mm / min corresponds to a tunneling score of 1.

[0104] Step S23: Use the analytic hierarchy process (AHP) to determine the weighting coefficients of the surrounding rock score and the tunneling score, then sum the weighted scores of the surrounding rock score and the tunneling score to calculate the rock-machine composite score.

[0105] More specifically, step S23 includes the following steps:

[0106] Step S23A: Use the analytic hierarchy process (AHP) to determine the weighting coefficients of the surrounding rock score and tunneling score for Class III surrounding rock, sum the weighted scores of the surrounding rock score and tunneling score, and calculate the rock-machine composite score to obtain the TBM tunneling score for Class III surrounding rock.

[0107] Step S23B: The weighting coefficients of the surrounding rock score and tunneling score for Class II and Class IV surrounding rock are determined by the analytic hierarchy process (AHP) respectively. The weighted sum of the surrounding rock score and tunneling score is calculated to obtain the rock-machine composite score for TBM tunneling in Class II or Class IV surrounding rock. In this step, the surrounding rock score for Class II and Class IV surrounding rock is the same, both being 3, so the same calculation method for the rock-machine composite score is used.

[0108] Step S23C: The weighting coefficients of the surrounding rock score and tunneling score for Class I and Class V surrounding rock are determined by the analytic hierarchy process (AHP). The surrounding rock score and tunneling score are weighted and summed to calculate the rock-machine composite score for TBM tunneling in Class I or Class V surrounding rock. In this step, the surrounding rock score for Class I and Class V surrounding rock is the same, which is 1. Therefore, the same calculation method for the rock-machine composite score is used.

[0109] In one embodiment, steps S23A, S23B, and S23C involve using the analytic hierarchy process (AHP) to determine the weighting coefficients of the surrounding rock score and tunneling score for each level of surrounding rock. This specifically includes the following steps:

[0110] Step S231: Construct the judgment matrix R based on the surrounding rock score s1 and the tunneling score s2 of this level of surrounding rock;

[0111] Step S232: Calculate the eigenvector v corresponding to the largest eigenvalue of the judgment matrix;

[0112] Step S233: Normalize the feature vector v to obtain the weight coefficients of the surrounding rock score and the tunneling score.

[0113] In one specific embodiment, step S23A specifically includes the following steps:

[0114] S23A1: The judgment matrix constructed from the surrounding rock score and tunneling score of Class III surrounding rock is as follows:

[0115]

[0116] Step S23A2: Calculate the eigenvector v = [0.7071, 0.7071] corresponding to the largest eigenvalue of the judgment matrix;

[0117] Step S23A3: Normalize the feature vector v to obtain the weight coefficients of the surrounding rock score and the tunneling score, which are 0.5000 and 0.5000, respectively;

[0118] Step S23A4: Calculate and obtain the rock-machine composite score CS based on the surrounding rock score s1, the tunneling score s2, and the weighting coefficient w2 of the surrounding rock score w1 and the tunneling score.

[0119] In one specific embodiment, step S23B specifically includes the following steps:

[0120] S23B1: The judgment matrix constructed from the scores of Class II and Class IV surrounding rock and the tunneling score is as follows:

[0121]

[0122] Step S23B2: Calculate the eigenvector v = [0.3162, 0.9487] corresponding to the largest eigenvalue of the judgment matrix.

[0123] Step S23B3: Normalize the feature vector v to obtain the weight coefficients of the surrounding rock score and the tunneling score, which are 0.2500 and 0.7500, respectively;

[0124] Step S23B4: Calculate and obtain the rock-machine composite score CS based on the surrounding rock score s1, the tunneling score s2, and the weighting coefficients w1 and w2 of the surrounding rock score and the tunneling score.

[0125] In one specific embodiment, step S23C specifically includes the following steps:

[0126] S23C1: The judgment matrix constructed from the surrounding rock scores and tunneling scores of Class I and Class V surrounding rock is as follows:

[0127]

[0128] Step S23C2: Calculate the eigenvector vv = [0.1961, 0.9806] corresponding to the largest eigenvalue of the judgment matrix;

[0129] Step S23C3: Normalize the feature vector v to obtain the weight coefficients of the surrounding rock score and the tunneling score, which are 0.1667 and 0.8333, respectively;

[0130] Step S23C4: Calculate and obtain the rock-machine composite score CS based on the surrounding rock score s1, the tunneling score s2, and the weighting coefficients w1 and w2 of the surrounding rock score and the tunneling score.

[0131] Steps S23A1, S23B1, and S23C1 determine the elements of the matrix, denoted as r. nm Where n is the number of rows, m is the number of columns, and r is the number of columns. nm Assign values ​​according to the following criteria:

[0132] Table 1 determines matrix element r nm Assignment criteria

[0133] <![CDATA[r nm ]]> Assignment Standard 1 Object n and object m are equally important 3 Objects n and m are slightly more important 5 Object n and object m are clearly more important 1 / 3 Objects n and m are slightly more important 1 / 5 Object n and object m are clearly more important

[0134] In Table 1, object n and object m are the surrounding rock score and tunneling score, respectively.

[0135] In one embodiment, step S24: Based on the rock-machine composite score, establish a grading standard for the surrounding rock at the TBM tunnel face; specifically, the surrounding rock grade within the range of [1, 2) is determined as Grade 1, the surrounding rock grade within the range of [2, 3) is determined as Grade 2, the surrounding rock grade within the range of [3, 4) is determined as Grade 3, and the surrounding rock grade within the range of [4, 5] is determined as Grade 4, as shown in Table 2. The higher the surrounding rock grade, the more conducive it is to TBM tunneling.

[0136] Table 2. Comparison of Rock Machine Composite Score and Surrounding Rock Grade at TBM Tunnel Face

[0137] Rock machine composite score TBM tunnel face surrounding rock grade [1，2) Level 1 [2，3) Level 2 [3，4) Level 3 [4，5] Level 4

[0138] In one embodiment, such as Figure 7 As shown, step S3: Based on the TBM tunnel face surrounding rock classification standard, constructing a database for training the rock-machine composite score cascade prediction model specifically includes the following steps:

[0139] Step S31: Collect multi-source information related to rock excavation from the excavated sections of multiple TBM projects; specifically, collect the on-site penetration index, torque penetration index, rock breaking energy, three-dimensional volume of rock debris, three-dimensional sphericity of rock debris, current conduction characteristics, and rock breaking dust characteristics collected in the aforementioned steps S1A-S1D, totaling seven dimensions of multi-source information related to rock excavation.

[0140] Step S32: Based on the constructed TBM tunnel face surrounding rock classification standard, the collected multi-source information related to the rock excavation machine during the TBM tunneling process is compared with the composite rock excavation machine score to obtain the composite rock excavation machine score.

[0141] Step S33: Using the collected multi-source information related to rock machines during the TBM tunneling process as input and the corresponding composite rock machine score as target output, construct a database including the input and target output for training the cascaded prediction model of rock machine scores.

[0142] In one embodiment, such as Figure 8 As shown, step S4: using the constructed database as training samples, constructing a rock machine composite score cascade prediction model based on ridge regression model and artificial neural network, specifically includes the following steps:

[0143] Step S41: Multiple training sets are obtained by sampling the database with replacement. The size of each training set is the same as the size of the database, i.e., the sample size is the same.

[0144] Step S42: Construct ridge regression models using multiple training sets as the first-level rock machine composite score prediction model;

[0145] Step S43: Input the training samples from the database into the first-level rock machine composite score prediction model to obtain the first-level prediction results;

[0146] Step S44: Combine the first-level prediction results and use them as the updated training samples for the input part while keeping the target output unchanged. Based on the updated database, use an artificial neural network algorithm to construct a second-level rock-machine composite score prediction model, and use the output of the second-level rock-machine composite score prediction model as the final prediction result. That is, use the second-level rock-machine composite score prediction model as the constructed rock-machine composite score prediction model.

[0147] In one specific embodiment, step S4: using the constructed database as training samples, constructing a rock machine composite score cascade prediction model based on the ridge regression model and artificial neural network, specifically includes the following steps:

[0148] The constructed database is sampled with replacement to obtain ten training sets, the size of which is the same as the size of the database.

[0149] A ridge regression model is constructed for each training set as the first-level rock machine composite score prediction model. The objective function f(w) of the ridge regression model is shown in the following equation:

[0150] f(w) = ||Xw - y||2 2 +α||w||2 2

[0151] In the formula, X refers to the input vector of the training sample; y refers to the target output of the training sample; ||·|2 refers to the Euclidean norm; w refers to the coefficients to be solved in the ridge regression model, which are obtained by minimizing f(w) using the least squares method. The least squares method for minimizing f(w) is shown in the following formula:

[0152]

[0153] The training samples from the database are input into each first-level rock-machine composite score prediction model to obtain multiple first-level prediction results. The set of multiple first-level prediction results is then used as the training samples for the new input part. in, The i-th ridge regression model predicts the results, while the target output y remains unchanged. An updated database is constructed based on the new input and the new target output.

[0154] An artificial neural network model is constructed based on the updated database as the second-level rock machine composite score prediction model, and its output is used as the final result. The neuron connection weights and biases in the artificial neural network model are obtained by minimizing Loss(weight, bias) using the error backpropagation method, as shown in the following equation:

[0155]

[0156] In one embodiment, such as Figure 9 As shown, step S5 includes the following steps:

[0157] Input information on the interaction between the rock excavator and the rock machine, the morphology of rock debris, the current conduction, and the rock-breaking dust.

[0158] The combined input is fed into the rock-machine composite score cascade prediction model to obtain the current rock-machine composite score of the working face;

[0159] The current rock grade of the tunnel face is obtained based on the established TBM tunnel surrounding rock classification standard.

[0160] In one specific embodiment, the real-time intelligent sensing method for the surrounding rock grade at the TBM tunnel face provided in this application further includes the following steps:

[0161] Step S6: Verify the predictive performance, i.e., the sensing accuracy, of the constructed rock-machine composite score cascade prediction model, and obtain the model performance verification results; specifically, this includes the following steps:

[0162] Get the test set;

[0163] Based on the test set, the prediction accuracy and F1-score of the rock machine composite score cascade prediction model are calculated and obtained to verify the model's prediction performance, i.e., the perception accuracy, and to obtain the model performance verification results.

[0164] In one specific embodiment, step S6 includes the following steps:

[0165] To verify the sensing effect of the real-time intelligent sensing method for the surrounding rock grade of TBM tunnel face in this application, a total of 120,576 samples were collected as a test set from a deep-buried TBM water conveyance tunnel project in my country. Among them, there were 26,261 samples of Class I surrounding rock, 32,765 samples of Class II surrounding rock, 32,657 samples of Class III surrounding rock, and 28,883 samples of Class IV surrounding rock, as shown in Table 3.

[0166] Table 3. Confusion matrix of the prediction results of the real-time intelligent sensing method for surrounding rock grade at the TBM tunnel face in this application for these 120,576 samples.

[0167]

[0168] To quantitatively evaluate the sensing accuracy of this application, two metrics are introduced: accuracy and F1-score. The former is a global metric used to measure the predictive performance of this application across the entire test set, calculated by Equation 10; the latter is a local metric used to measure the predictive performance of this application for each type of surrounding rock in the test set, calculated by Equation 11. Evaluation results show that the predictive accuracy of this application is 97.88%, and the F1-scores for Class I, Class II, Class III, and Class IV surrounding rock are 0.9774, 0.9796, 0.9794, and 0.9786, respectively.

[0169]

[0170] In the formula, K refers to the number of sample categories; n ij This refers to the number of samples where the true class is i and the predicted class is j.

[0171] This paper compares and analyzes the perception performance using multi-source information related to the rock machine with that using only TBM operating parameters as input information. Table 4 shows the confusion matrix of the perception model's prediction results for these 120,576 samples when only TBM operating parameters are used as input information. The prediction accuracy is 61.81%, and the F1-scores for Class I, Class II, Class III, and Class IV surrounding rock are 0.5651, 0.6461, 0.6473, and 0.6012, respectively.

[0172] Table 4 shows the confusion matrix of the perceptual model's prediction results for these 120,576 samples when only the TBM running parameters are used as part of the input information.

[0173]

[0174] In comparison, the prediction accuracy of the perception model using multi-source information related to rock machinery in this application increased by 36.07%, and the F1-scores for Class I, Class II, Class III, and Class IV surrounding rock were improved by 0.4123, 0.3335, 0.3322, and 0.3775, respectively. Figure 10 As shown.

[0175] Secondly, this application provides a real-time intelligent sensing system for the surrounding rock grade at the tunnel face during TBM excavation, comprising:

[0176] The multi-source information acquisition module is used to collect and acquire multi-source information related to the rock machine during the TBM tunneling process;

[0177] The grading standard construction module is used to establish grading standards for the surrounding rock of TBM tunnel faces;

[0178] The database construction module is communicatively connected to the grading standard construction module and is used to construct a database for training the rock machine composite score cascade prediction model based on the TBM tunnel face surrounding rock grading standard.

[0179] The prediction model building module communicates with the database building module and is used to build a rock machine composite score cascade prediction model based on the ridge regression model and artificial neural network, using the built database as training samples.

[0180] The tunnel face surrounding rock grade acquisition module communicates with the multi-source information acquisition module, the prediction model construction module, and the grading standard construction module. It is used to input the multi-source information related to rock machinery during the TBM tunneling process into the constructed rock machinery composite score cascade prediction model to obtain the composite rock machinery score of the current TBM tunnel face. Then, based on the constructed TBM tunnel face surrounding rock grading standard, the TBM tunnel face surrounding rock grade is obtained.

[0181] In conjunction with the second aspect, in one implementation, the multi-source information acquisition module includes:

[0182] The rock-machine interaction information acquisition unit is used to collect rock-machine interaction information during the TBM tunneling process;

[0183] Rock debris morphology information acquisition unit, used to collect rock debris morphology information during TBM tunneling;

[0184] The current conduction information acquisition unit is used to collect current conduction information during the TBM tunneling process;

[0185] The rock-breaking dust information acquisition unit is used to collect rock-breaking dust information during the TBM tunneling process. The functions of each module in the aforementioned TBM tunnel face surrounding rock grade real-time intelligent sensing system correspond to the steps in the aforementioned TBM tunnel face surrounding rock grade real-time intelligent sensing method embodiment; their functions and implementation processes will not be elaborated upon here.

[0186] Thirdly, this application provides a real-time intelligent sensing device for the surrounding rock grade of a TBM tunnel face during excavation. The real-time intelligent sensing device for the surrounding rock grade of a TBM tunnel face during excavation can be a personal computer (PC), a laptop computer, a server, or other devices with data processing capabilities.

[0187] The communication interface includes input / output (I / O) interfaces, physical interfaces, and logical interfaces. These interfaces enable interconnection of internal components within the TBM tunnel face surrounding rock grade real-time intelligent sensing device, as well as interfaces for interconnection between the device and other devices (such as other computing devices or user equipment). Physical interfaces can be Ethernet interfaces, fiber optic interfaces, ATM interfaces, etc.; user equipment can be displays, keyboards, etc.

[0188] Memory can be various types of storage media, such as random access memory (RAM), read-only memory (ROM), non-volatile RAM (NVRAM), flash memory, optical storage, hard disk, programmable ROM (PROM), erasable PROM (EPROM), electrically erasable PROM (EEPROM), etc.

[0189] The processor can be a general-purpose processor, which can call the real-time intelligent sensing program for the surrounding rock grade of the TBM tunnel face stored in the memory and execute the real-time intelligent sensing method for the surrounding rock grade of the TBM tunnel face provided in the embodiments of this application. For example, the general-purpose processor can be a central processing unit (CPU). The method executed when the real-time intelligent sensing program for the surrounding rock grade of the TBM tunnel face is called can refer to the various embodiments of the real-time intelligent sensing method for the surrounding rock grade of the TBM tunnel face in this application, and will not be repeated here.

[0190] Fourthly, embodiments of this application also provide a readable storage medium.

[0191] The present application stores a real-time intelligent sensing program for the surrounding rock grade of a TBM tunnel face on a readable storage medium. When the TBM tunnel face surrounding rock grade real-time intelligent sensing program is executed by a processor, it implements the steps of the TBM tunnel face surrounding rock grade real-time intelligent sensing method as described above.

[0192] The method implemented when the TBM tunnel face surrounding rock grade real-time intelligent sensing program is executed can be referred to in various embodiments of the TBM tunnel face surrounding rock grade real-time intelligent sensing method of this application, and will not be repeated here.

[0193] It should be noted that the sequence numbers of the embodiments in this application are for descriptive purposes only and do not represent the superiority or inferiority of the embodiments.

[0194] Through the above description of the embodiments, those skilled in the art can clearly understand that the methods of the above embodiments can be implemented by means of software plus necessary general-purpose hardware platforms. Of course, they can also be implemented by hardware, but in many cases the former is a better implementation method. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product is stored in a storage medium (such as ROM / RAM, magnetic disk, optical disk) as described above, and includes several instructions to cause a terminal device to execute the methods described in the various embodiments of this application.

[0195] The above are merely preferred embodiments of this application and do not limit the patent scope of this application. Any equivalent structural or procedural transformations made using the content of this application's specification and drawings, or direct or indirect applications in other related technical fields, are similarly included within the patent protection scope of this application.

Claims

1. A method for real-time intelligent sensing of the surrounding rock grade at the tunnel face during TBM excavation, characterized in that, Includes the following steps: Collect and acquire multi-source information related to the rock machine during the TBM tunneling process. The multi-source information related to the rock machine includes multi-source information related to the TBM and the surrounding rock. Establish a classification standard for the surrounding rock at the tunnel face of a TBM, including: Determine the score for the surrounding rock; Determine the tunneling score; The weighting coefficients of the surrounding rock score and the tunneling score are determined by the analytic hierarchy process (AHP), and the weighted sum of the surrounding rock score and the tunneling score is used to calculate the rock-machine composite score. A classification standard for surrounding rock at the tunnel face of a TBM is established based on the rock-machine composite score. Based on the TBM tunnel face surrounding rock classification standard, a database is constructed for training the rock machine composite score cascade prediction model; Using the constructed database as training samples, a cascaded prediction model for rock machine composite scores is built based on ridge regression and artificial neural networks, including: Multiple training sets are obtained by sampling the database with replacement, wherein the size of the training sets is the same as the size of the database. Ridge regression models were constructed using multiple training sets to serve as the first-level rock machine composite score prediction model. Input the training samples from the database into the first-level rock machine composite score prediction model to obtain the first-level prediction results; The first-level prediction results are combined and used as the input of the updated training samples while keeping the target output unchanged. An artificial neural network model is constructed based on the updated database to build a second-level rock machine composite score prediction model. The output of the second-level rock machine composite score prediction model is used as the final prediction result. Input the multi-source information related to rock machinery during the TBM tunneling process into the constructed rock machinery composite score cascade prediction model to obtain the composite rock machinery score of the current TBM tunnel face. Then, based on the constructed TBM tunnel face surrounding rock classification standard, obtain the TBM tunnel face surrounding rock grade.

2. The real-time intelligent sensing method for the surrounding rock grade at the TBM tunnel face as described in claim 1, characterized in that, The acquisition of multi-source information related to the rock excavator during the TBM tunneling process specifically includes the following steps: Collect information on the interaction between the TBM and the rock excavator during the tunneling process; Collect information on the morphology of rock debris during TBM tunneling; Collect current conduction information during the TBM tunneling process; Collect information on rock-breaking dust during the TBM tunneling process.

3. The real-time intelligent sensing method for the surrounding rock grade at the TBM tunnel face as described in claim 1, characterized in that, The database for training the rock-machine composite score cascade prediction model, based on the TBM tunnel face surrounding rock classification standard, specifically includes the following steps: Collect multi-source information related to rock excavation machines during the TBM tunneling process from excavated sections of multiple TBM projects; The collected multi-source information related to rock machines during the TBM tunneling process is compared with the constructed TBM tunnel face surrounding rock classification standard to obtain a composite rock machine score. Using the collected multi-source information related to rock machines during the TBM tunneling process as input and the corresponding composite rock machine score as target output, a database including the input and target output is constructed for training a cascaded prediction model of rock machine scores.

4. The real-time intelligent sensing method for the surrounding rock grade at the TBM tunnel face as described in claim 1, characterized in that, In the step of constructing ridge regression models using multiple training sets as the first-level rock-machine composite score prediction model, the objective function of the ridge regression model is... As shown in the following formula: In the formula, The input vector of the training samples; The target output of the training samples; The Euclidean norm; The coefficients to be solved in the Finger Ridge regression model.

5. The real-time intelligent sensing method for the surrounding rock grade at the TBM tunnel face as described in claim 1, characterized in that, The loss function of the artificial neural network model As shown in the following formula: In the formula, and These represent the neuron connection weights and biases in an artificial neural network model.

6. The real-time intelligent sensing method for the surrounding rock grade at the TBM tunnel face as described in claim 1, characterized in that, It also includes the following steps: Get the test set; Based on the test set, the perception performance of the rock machine composite score prediction model is calculated and the model's perception performance is verified.

7. A sensing system based on the real-time intelligent sensing method for the surrounding rock grade of a TBM tunnel face as described in any one of claims 1-6, characterized in that, The system includes: The multi-source information acquisition module is used to collect and acquire multi-source information related to the rock machine during the TBM tunneling process; The grading standard construction module is used to establish grading standards for the surrounding rock of TBM tunnel faces; The database construction module is communicatively connected to the grading standard construction module and is used to construct a database for training the rock machine composite score cascade prediction model based on the TBM tunnel face surrounding rock grading standard. The prediction model building module communicates with the database building module and is used to build a rock machine composite score cascade prediction model based on the ridge regression model and artificial neural network, using the built database as training samples. The tunnel face surrounding rock grade acquisition module communicates with the multi-source information acquisition module, the prediction model construction module, and the grading standard construction module. It is used to input the multi-source information related to rock machinery during the TBM tunneling process into the constructed rock machinery composite score cascade prediction model to obtain the composite rock machinery score of the current TBM tunnel face. Then, based on the constructed TBM tunnel face surrounding rock grading standard, the TBM tunnel face surrounding rock grade is obtained.

8. The real-time intelligent sensing system for the surrounding rock grade of a TBM tunnel face as described in claim 7, characterized in that, The multi-source information acquisition module includes: The rock-machine interaction information acquisition unit is used to collect rock-machine interaction information during the TBM tunneling process; Rock debris morphology information acquisition unit, used to collect rock debris morphology information during TBM tunneling; The current conduction information acquisition unit is used to collect current conduction information during the TBM tunneling process; The rock-breaking dust information acquisition unit is used to collect rock-breaking dust information during the TBM tunneling process.

Images