TBM intelligent decision-making system and method based on rock quartz content and cutterhead abrasion

Abstract

According to the TBM intelligent decision-making system and method based on the rock quartz content and the cutter head abrasion, the optimal tunneling parameters can be intelligently selected based on the rock quartz content data and the TBM parameters, the abrasion of the hobbing cutter during construction of the full-section rock tunnel boring machine can be effectively controlled and reduced, and the abrasion loss of the hobbing cutter can be accurately predicted. At present, the selection of tunneling parameters is mostly based on the maximum power during TBM steady-state tunneling, and according to the method, the service life of a disc cutter is greatly prolonged, the cutter replacement time is shortened and the engineering management level is improved on the basis of achieving TBM intelligent tunneling by combining the maximum power and reasonably considering cutter abrasion.

Description

Technical Field

[0001] The present invention belongs to the technical field of intelligent control of tunnel boring equipment construction, and specifically relates to a TBM intelligent decision-making system and method based on rock quartz content and cutterhead wear. Background Art

[0002] The statements in this section merely provide background information related to the present disclosure and do not necessarily constitute prior art.

[0003] Disc cutters are the core rock-breaking tools of full-face rock tunnel boring machines (TBMs). Cutter wear has become one of the key problems that are difficult to avoid during TBM construction. Cutter wear is mainly controlled by geological factors and mechanical factors. The main reason for rock wear on the cutter is the quartz content of the rock, and the mechanical reason is the design parameters of the cutter and the selection of operating parameters such as thrust, cutter torque and penetration during TBM excavation.

[0004] Currently, in the traditional construction process, there is a lack of consideration for the wear of the cutterhead. At the same time, the selection of operating parameters mainly depends on the experience of the driver in the TBM main control room, and the various parameters of TBM excavation are selected by observing the data from sensors during the advancement process.

[0005] The above methods have many disadvantages:

[0006] (1) Because the cutterhead wear problem is not taken into consideration, it is difficult to predict and quickly adjust the operating parameters of the TBM in a complex geological environment, which will cause severe wear of the disc cutter and affect the service life of the cutterhead.

[0007] (2) The experience required of the driver in the TBM main control room is relatively high, and it is difficult to ensure the correctness of the technician's operation every time, which may cause wear and tear on the cutterhead due to abnormal operation.

[0008] (3) Personnel are required to continuously debug the operating parameters until the TBM can operate stably. This process is inefficient and will waste a lot of construction time. Summary of the invention

[0009] In order to solve the above problems, the present invention proposes a TBM intelligent decision-making system and method based on rock quartz content and cutter disc wear. The present invention considers the quartz content and combines it to realize automatic TBM optimal operating parameter selection, extend the life of the disc cutter, and improve the excavation speed.

[0010] According to some embodiments, the present disclosure adopts the following technical solutions:

[0011] A TBM intelligent decision-making method based on rock quartz content and cutterhead wear includes the following steps:

[0012] Step 1: When the TBM is working, the rock powder is tested to obtain a spectrum, and the spectrum is analyzed to obtain the rock quartz content data, and obtain various TBM excavation parameters, including cutterhead speed, cutterhead torque, excavation speed, total propulsion force and penetration;

[0013] Step 2: Cache and package the collected data in segments, and store the rock quartz content data and TBM excavation parameters in different databases;

[0014] Step 3: Establish a database at the terminal to store the rock quartz content and TBM ascending section excavation parameters respectively;

[0015] Step 4: Filter abnormal data to extract correct parameters. Abnormal data includes abnormal working status data, startup status data and manual operation error data;

[0016] Step 5: Take the TBM excavation parameters of the ascending section as the input layer data for deep learning to establish a rock machine data training model and use the state parameters of the surrounding rock in the stable section as the output layer for prediction to obtain the uniaxial compressive strength of the rock U and the number of joints per unit volume of the rock mass J. v , surrounding rock grade W;

[0017] Step 6: Establish the state parameter matrix of the surrounding rock in the stable section and use the support vector machine algorithm model to predict the steady-state section excavation parameters of the TBM considering only the maximum excavation power. The kernel function type of the support vector machine algorithm model is the radial kernel function. The cross-validation method is used to search for the optimal penalty factor and kernel function parameters of the model, and the sequence minimization algorithm is used to train the support vector machine algorithm model.

[0018] Step 7: Establish the cutter field cutting depth index and the average friction rate of the positive cutter according to the TBM excavation parameters of the ascending section;

[0019] Step 8: Take the average wear rate of the positive cutter in the TBM ascending section as the input layer for deep learning, establish a data training model and use the various TBM parameters as the output layer for prediction, and obtain the TBM stable section excavation parameters considering the cutter wear;

[0020] Step 9: Establish rock wear resistance index and rock abrasiveness index according to the state parameters of the surrounding rock in the stable section and the quartz content of the rock, and classify the friction of the surrounding rock according to the rock abrasiveness index;

[0021] Step 10: Obtain the optimal excavation parameters considering the cutterhead wear based on the classification results, output them to the TBM client in real time, and issue instructions to adjust various TBM excavation parameters at the same time.

[0022] As an alternative embodiment, rock quartz content data and TBM cutterhead operating parameters are stored in different databases.

[0023] As an optional implementation plan, a support vector machine algorithm model is used to conduct deep learning on the TBM cutterhead operating parameters to establish a rock-machine data feedback model and use the stable section surrounding rock state parameters as the output layer matrix, including rock compressive strength, number of joints per unit volume of rock mass and surrounding rock grade.

[0024] As an optional implementation scheme, the cutter field depth index FPI = F / P·N, where F represents the cutter disc thrust, P represents the cutter disc penetration, and N represents the number of cutters on the cutter disc, thereby obtaining the positive cutter average friction rate Y1 = 0.0025lnX1-0.0929, where Y1 represents the positive cutter average wear rate, and X1 represents the field depth index FPI.

[0025] As an optional implementation scheme, the rock wear resistance index RAI = U × EQC, where U represents the uniaxial compressive strength of the rock, and EQC represents the equivalent quartz content of the rock, thereby obtaining the rock abrasiveness index CAI:

[0026] Y2=0.996lnX2-0.656, Y2 represents the rock abrasiveness index CAI, and X2 represents the rock wear resistance index RAI.

[0027] As an optional implementation scheme, the specific process of classifying according to the rock abrasiveness index includes:

[0028] (1) When CAI < 0.3, the surrounding rock is considered to have no friction. When CAI is between 0.3 and 0.5, the surrounding rock is considered to have low friction. In this case, there is no need to consider the wear of the cutterhead for parameter adjustment.

[0029] (2) When CAI is between 0.5 and 1.0, the surrounding rock is considered to have medium friction. When CAI is between 1.0 and 2.0, the surrounding rock is considered to have friction. At this time, the operating parameters should be fine-tuned, and the parameter weight considering cutterhead wear is 30%;

[0030] (3) When CAI is between 2.0 and 4.0, the surrounding rock is considered to have high friction. When CAI is between 4.0 and 6.0, the surrounding rock is considered to have extremely high friction. At this time, the operating parameters should be adjusted, and the parameter weight of the cutterhead wear is considered to be 50%.

[0031] As an optional implementation method, if the predicted value of the excavation parameter fluctuates greatly within a short period of time, the system will sound an alarm, and the driver will judge the on-site situation and manually control and take corresponding measures.

[0032] A TBM intelligent decision-making system based on rock quartz content and cutterhead wear, including:

[0033] The parameter acquisition module is configured to detect rock powder to obtain a spectrum when the TBM is working, analyze the spectrum to obtain rock quartz content data, and obtain TBM cutterhead operation parameters;

[0034] The data preprocessing module is configured to cache and package the collected data in segments; filter abnormal data to extract the correct parameters of the TBM ascent segment as input layer data;

[0035] The modeling module is configured to use the support vector machine algorithm model to perform deep learning on the input layer data, so as to establish a rock-machine data feedback model and use the state parameters of the surrounding rock in the stable section as the output layer for prediction; establish a rock wear resistance index based on the state parameters of the surrounding rock in the stable section and the quartz content of the rock, and then obtain the wear degree of the cutter; establish a matrix of the state parameters of the surrounding rock in the stable section and consider the wear degree of the cutter head, and use the support vector machine algorithm model to predict the excavation parameters of the steady-state section;

[0036] The parameter adjustment module is configured to perform weighted processing on the TBM excavation parameters according to the rock friction classification, and output them to the upper computer visualization interface in real time. At the same time, the system sends instructions to the PLC controller to adjust to the optimal excavation parameters of the TBM stable section;

[0037] The model update module is configured to periodically update the model according to the defined time and upgrade the training model of the input layer data.

[0038] A computer-readable storage medium stores a plurality of instructions, wherein the instructions are suitable for being loaded by a processor of a terminal device and executing part or all of the steps of a TBM intelligent decision-making method based on rock quartz content and cutterhead wear.

[0039] A terminal device includes a processor and a computer-readable storage medium, wherein the processor is used to implement various instructions; and the computer-readable storage medium is used to store multiple instructions, wherein the instructions are suitable for being loaded by the processor and executing part or all of the steps of a TBM intelligent decision-making method based on rock quartz content and cutterhead wear.

[0040] Compared with the prior art, the present invention has the following beneficial effects:

[0041] The present invention can intelligently select the best excavation parameters based on the rock quartz content data and TBM parameters, effectively control and reduce the wear of the disc cutter during the construction of the full-section rock tunnel boring machine, and accurately predict the wear of the disc cutter. At present, the selection of excavation parameters is mostly based on the maximum power of the TBM during steady-state excavation. The present invention combines the maximum power and reasonably considers the wear of the cutter head, which will greatly extend the life of the disc cutter on the basis of realizing TBM intelligent excavation, reduce the time of replacing the cutter head, and improve the level of engineering management. BRIEF DESCRIPTION OF THE DRAWINGS

[0042] The accompanying drawings constituting a part of the present disclosure are used to provide a further understanding of the present disclosure. The illustrative embodiments of the present disclosure and their descriptions are used to explain the present disclosure and do not constitute an improper limitation on the present disclosure.

[0043] Figure 1 This is the TBM excavation intelligent decision-making flow chart disclosed in this disclosure;

[0044] Figure 2 This is a structural diagram of the TBM excavation intelligent control system disclosed in the present invention;

[0045] Figure 3 This is a schematic diagram of the data repository disclosed in the present invention;

[0046] Figure 4 A flowchart is obtained for the intelligent decision-making of the present disclosure. DETAILED DESCRIPTION

[0047] The present disclosure is further described below in conjunction with the accompanying drawings and embodiments.

[0048] It should be noted that the following detailed descriptions are all illustrative and are intended to provide further explanation of the present disclosure. Unless otherwise specified, all technical and scientific terms used herein have the same meanings as those commonly understood by those skilled in the art to which the present disclosure belongs.

[0049] It should be noted that the terms used herein are only for describing specific embodiments and are not intended to limit the exemplary embodiments according to the present disclosure. As used herein, unless the context clearly indicates otherwise, the singular form is also intended to include the plural form. In addition, it should be understood that when the terms "comprising" and / or "including" are used in this specification, it indicates the presence of features, steps, operations, devices, components and / or combinations thereof.

[0050] An intelligent decision-making system for TBM based on rock quartz content and cutterhead wear, such as Figure 1 As shown, it includes an acquisition module, a data storage module, an abnormal data filtering module, a data perception feedback module, an intelligent decision-making module, a prediction data output and automation control module and a self-learning update module.

[0051] The data acquisition module is configured to collect rock quartz content data; the second is configured to obtain various performance parameters of TBM excavation;

[0052] The background data storage module is used to store rock quartz content data and TBM excavation parameters and store them in different data tables;

[0053] The abnormal data filtering module filters the abnormal data to extract the correct parameters of the TBM ascending section;

[0054] The data perception feedback module uses a support vector machine algorithm to perform deep learning on the input layer data, and establishes a rock machine data feedback model to predict the state of the surrounding rock in the stable section;

[0055] The intelligent decision-making module estimates the TBM performance parameters according to the obtained surrounding rock state parameters;

[0056] The forecast data output and automatic control module, firstly, sends instructions to the PLC controller to adjust the corresponding parameters of the TBM to achieve full automatic control; secondly, it provides real-time feedback on the forecast data, and can manually adjust according to the forecast data in special circumstances;

[0057] The self-learning update module updates the model at a manually defined time to upgrade the training model of the input layer data.

[0058] The specific work process includes:

[0059] Step 1: The rock powder is tested by XRD to obtain a spectrum, and five parameters including TBM cutter head speed n, cutter head torque T, excavation speed V, total propulsion force F, and penetration P are collected according to the sensors of various TBM equipment components;

[0060] Step 2: If Figure 2 As shown, the spectrum and various TBM parameters are transmitted to the TBM host computer client through the PLC data transmission interface, and the spectrum is analyzed by software to obtain the rock quartz content data;

[0061] Step 3: OPC standard access PLC caches the collected data in the local database, and then periodically packages the cached data in segments and uploads them to the background data storage module;

[0062] Step 4: Establish a large data repository on the backend server. The storage steps are as follows according to the different data storage types. Figure 3 As shown:

[0063] Step 1: Establish data prediction system and model update system in the background;

[0064] Step 2: Sub-library A and sub-library B are established in the data prediction system to store rock quartz content data and various excavation parameter data of the ascending section TBM respectively, which are used to predict the five parameters of TBM cutterhead speed n, cutterhead torque T, excavation speed V, total propulsion force F, and penetration P during the stable section excavation;

[0065] Step 3: Establish sub-library C in the model updating system to store various excavation parameters of the TBM stable section for self-learning and updating the prediction model.

[0066] Step 5: To avoid abnormal data and data under some special excavation conditions, which may cause low-quality machine learning results due to the presence of noise, incompleteness, and chaos in the original data in the machine learning database, the data is preprocessed. The specific preprocessing steps are as follows:

[0067] Step 1: Data cleaning. For data with noise in the database, the data is mainly processed by binning. The data to be processed is classified and examined, and each type of data is smoothed. For the erroneous data in the database, it is changed to the average value of several nearby data.

[0068] Step 2: Perform data integration, remove redundant data, and then merge data from multiple data sources and store them in a database or file;

[0069] Step 3: Perform data transformation, which mainly normalizes the data and converts the data into a form that is easy to mine, that is, the number of attributes is scaled so that it falls into a small specific interval. This embodiment uses the minimum-maximum normalization method to perform linear transformation on the data. Assume that the maximum and minimum values ​​of attribute A are max and A and min A , by calculating:

[0070]

[0071] Map an original value v in A to an interval through min-max normalization The value v′ in the minimum-maximum normalization can maintain the relationship between the original data. is [0, 1], that is, all data is mapped to [0, 1];

[0072] Step 4: Use the CfsSubsetEval method to reduce data. Without affecting the mining effect, the data is compressed by eliminating redundant attributes, numerical aggregation, etc.

[0073] Step 5: Use characteristic equation to filter abnormal data, including abnormal working status data, startup status data, and manual operation error data.

[0074] Step 6: If Figure 4 As shown in the figure, the support vector machine algorithm model is used to perform deep learning on the input layer data M1 = [n, T, V, F, P] to establish a rock machine data feedback model and use the state parameters of the surrounding rock in the stable section as the output layer matrix Y1 = [U, J v ,W] for prediction, where U represents the rock compressive strength U, J vrepresents the number of joints per unit volume of rock mass, W represents the surrounding rock grade, the independent variable x is the TBM performance parameter, and the dependent variable y is the surrounding rock state coefficient;

[0075] Based on V C The support vector machine algorithm based on the dimensional theory and the principle of result risk minimization is expressed using the radial basis kernel function. in is the kernel function of the support vector machine, which represents the nonlinear mapping from the input control to the high-dimensional feature space. ω1 is the weight vector, b1 is the bias term, and the Lagrange multiplier α=(α1,α2…α n ) T ∈R n ;

[0076] Its support vector expression is as follows:

[0077]

[0078] Select the radial kernel function RBF as the kernel function type, and its expression is as follows:

[0079]

[0080] in

[0081] The cross-validation method is used to search for the optimal penalty factor C and kernel function parameter g of the model. The upper limit of the regression error ε is set to 0.1. If the error is less than this value, no penalty is imposed, and if it is greater than this value, a penalty is imposed. The SVM model is trained using the sequential minimization algorithm.

[0082] Step 7: Establish the stable section surrounding rock state parameter matrix Y2 = [U, J v , W], using the support vector machine algorithm model to analyze the parameters of the TBM steady-state section considering only the maximum excavation power. Make a prediction, where the number of cutter head revolutions Cutter torque Advance speed Total propulsion force Penetration The independent variable x is the state parameter of the surrounding rock in the stable section, and the dependent variable y is the predicted parameter under the maximum excavation power of the TBM;

[0083] Based on V C The support vector machine algorithm based on the dimensional theory and the principle of result risk minimization is expressed using the radial basis kernel function. in is the kernel function of the support vector machine, which represents the nonlinear mapping from the input space to the high-dimensional feature space, ω2 is the weight vector, b2 is the bias term, and the Lagrange multiplier α = (α1, α2…α n )T ∈R n .

[0084] Its support vector expression is as follows:

[0085]

[0086] Select the radial kernel function RBF as the kernel function type, and its expression is as follows:

[0087]

[0088] in

[0089] The cross-validation method is used to search for the optimal penalty factor C and kernel function parameter g of the model. The upper limit of the regression error ε is set to 0.1. If the error is less than this value, no penalty is imposed, and if it is greater than this value, a penalty is imposed. The SVM model is trained using the sequential minimization algorithm.

[0090] Step 8: According to the TBM operation parameters of the ascending section, the cutter field cutting depth index FPI = F / P·N is established, where F represents the cutter head thrust, P represents the cutter head penetration, and N represents the number of cutters on the cutter head, and then the average friction rate of the positive cutter is obtained, where:

[0091] Y1=0.0025lnX1-0.0929, Y1 represents the average wear rate of the positive hob, and X1 represents the field depth of cut index FPI;

[0092] Step 9: Take the average wear rate of the positive disc cutter in the TBM ascending stage as the input layer for deep learning, and use the support vector machine algorithm model to calculate the parameters of the TBM steady-state section considering the cutterhead wear. Make a prediction, where the number of cutter head revolutions Cutter torque Advance speed Total propulsion force Penetration The independent variable x is the state parameter of the surrounding rock in the stable section, and the dependent variable y is the TBM prediction parameter considering the cutterhead wear;

[0093] Based on V C The support vector machine algorithm based on the dimensional theory and the principle of result risk minimization is expressed using the radial basis kernel function. in is the kernel function of the support vector machine, which represents the nonlinear mapping from the input space to the high-dimensional feature space, ω3 is the weight vector, b3 is the bias term, and the Lagrange multiplier α=(α1,α2…α n ) T ∈R n .

[0094] Its support vector expression is as follows:

[0095]

[0096] Select the radial kernel function RBF as the kernel function type, and its expression is as follows:

[0097]

[0098] in

[0099] The cross-validation method is used to search for the optimal penalty factor C and kernel function parameter g of the model. The upper limit of the regression error ε is set to 0.1. If the error is less than this value, no penalty is imposed, and if it is greater than this value, a penalty is imposed. The SVM model is trained using the sequential minimization algorithm.

[0100] Step 10: Establish the rock wear resistance index RAI=U×EQC according to the state parameters of the surrounding rock in the stable section and the quartz content of the rock, where U represents the uniaxial compressive strength of the rock and EQC represents the equivalent quartz content of the rock, and then obtain the rock abrasiveness index CAI, and classify the friction of the surrounding rock according to the rock abrasiveness index CAI;

[0101] Among them, Y2 = 0.996lnX2-0.656, Y2 represents the rock abrasiveness index CAI, X2 represents the rock wear resistance index RAI, and the surrounding rock is classified according to CAI as follows:

[0102] (1) When CAI < 0.3, the surrounding rock is considered to have no friction. When CAI is between 0.3 and 0.5, the surrounding rock is considered to have low friction. In this case, there is no need to consider the wear of the cutterhead for parameter adjustment.

[0103] (2) When CAI is between 0.5 and 1.0, the surrounding rock is considered to have medium friction. When CAI is between 1.0 and 2.0, the surrounding rock is considered to have friction. At this time, the operating parameters should be fine-tuned, and the parameter weight considering cutterhead wear is 30%;

[0104] (3) When CAI is between 2.0 and 4.0, the surrounding rock is considered to have high friction. When CAI is between 4.0 and 6.0, the surrounding rock is considered to have extremely high friction. At this time, the operating parameters should be adjusted, and the parameter weight considering cutterhead wear is 50%;

[0105] Step 11: Based on the CAI classification results, the excavation parameters that take into account the cutterhead wear are obtained, output to the TBM client in real time and displayed on the upper computer visualization interface; at the same time, the system sends instructions to the PLC controller to adjust to the optimal TBM excavation parameters. If the predicted value of the excavation parameter fluctuates greatly in a short period of time, the system will issue an alarm, and the driver will judge the on-site situation and manually control to take corresponding measures.

[0106] Step 12: The TBM excavation parameters of the stable section are collected in real time and stored in the data storage module as a sub-library C, and transmitted to the self-learning update module as a difference analysis and configured as an optimized training model.

[0107] The data repository of this embodiment can be continuously enriched. By using machine learning technology and a support vector machine method, a rock machine data training model is created to estimate the optimal excavation parameters of the TBM stable section under the condition of cutterhead wear. The prediction results can be displayed in real time on the upper computer operation interface, thereby realizing fully automated tunnel excavation under the condition of low cutterhead wear.

[0108] The model updating unit can self-learn the rock machine data training model according to the continuous updating of the data repository to adapt to TBMs with different surrounding rocks, different diameters and different performances.

[0109] Those skilled in the art will appreciate that the embodiments of the present disclosure may be provided as methods, systems, or computer program products. Therefore, the present disclosure may take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Moreover, the present disclosure may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0110] The present disclosure is described with reference to the flowcharts and / or block diagrams of the methods, devices (systems), and computer program products according to the embodiments of the present disclosure. It should be understood that each process and / or box in the flowchart and / or block diagram, as well as the combination of the processes and / or boxes in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0111] These computer program instructions may also be stored in a computer readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer readable memory produce an article of manufacture including an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0112] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0113] The above description is only a preferred embodiment of the present disclosure and is not intended to limit the present disclosure. For those skilled in the art, the present disclosure may have various modifications and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present disclosure shall be included in the protection scope of the present disclosure.

[0114] Although the above describes the specific implementation methods of the present disclosure in conjunction with the accompanying drawings, it is not intended to limit the scope of protection of the present disclosure. Technical personnel in the relevant field should understand that on the basis of the technical solution of the present disclosure, various modifications or variations that can be made by those skilled in the art without creative work are still within the scope of protection of the present disclosure.

Claims

1. A TBM intelligent decision-making method based on rock quartz content and cutterhead wear, characterized by: The following steps are involved: Step 1: When the TBM is working, the rock powder is tested to obtain a spectrum, and the spectrum is analyzed to obtain the rock quartz content data, and obtain various TBM excavation parameters, including cutterhead speed, cutterhead torque, excavation speed, total propulsion force and penetration; Step 2: Cache and package the collected data in segments, and store the rock quartz content data and TBM excavation parameters in different databases; Step 3: Establish a database at the terminal to store the rock quartz content and TBM ascending section excavation parameters respectively; Step 4: Filter abnormal data to extract correct parameters. Abnormal data includes abnormal working status data, startup status data and manual operation error data; Step 5: Take the TBM excavation parameters of the ascending section as the input layer data for deep learning to establish a rock machine data training model and use the state parameters of the surrounding rock in the stable section as the output layer for prediction to obtain the uniaxial compressive strength of the rock U and the number of joints per unit volume of the rock mass J. v , surrounding rock grade W; Step 6: Establish the state parameter matrix of the surrounding rock in the stable section and use the support vector machine algorithm model to predict the steady-state section excavation parameters of the TBM considering only the maximum excavation power. The kernel function type of the support vector machine algorithm model is the radial kernel function. The cross-validation method is used to search for the optimal penalty factor and kernel function parameters of the model, and the sequence minimization algorithm is used to train the support vector machine algorithm model. Step 7: Establish the cutter field cutting depth index and the average friction rate of the positive cutter according to the TBM excavation parameters of the ascending section; Step 8: Take the average wear rate of the positive cutter in the TBM ascending section as the input layer for deep learning, establish a data training model and use the various TBM parameters as the output layer for prediction, and obtain the TBM stable section excavation parameters considering the cutter wear; Step 9: Establish rock wear resistance index and rock abrasiveness index according to the state parameters of the surrounding rock in the stable section and the quartz content of the rock, and classify the friction of the surrounding rock according to the rock abrasiveness index; Step 10: Obtain the optimal excavation parameters considering the cutterhead wear based on the classification results, output them to the TBM client in real time, and issue instructions to adjust various TBM excavation parameters at the same time.

2. The TBM intelligent decision-making method based on rock quartz content and cutterhead wear according to claim 1 is characterized by: The rock quartz content data and TBM cutterhead operating parameters are stored in different databases.

3. The TBM intelligent decision-making method based on rock quartz content and cutterhead wear according to claim 1 is characterized by: The support vector machine algorithm model is used to conduct deep learning on the TBM cutterhead operating parameters to establish a rock-machine data feedback model and use the stable section surrounding rock state parameters as the output layer matrix, including rock compressive strength, number of joints per unit volume of rock mass and surrounding rock grade.

4. The TBM intelligent decision-making method based on rock quartz content and cutterhead wear according to claim 1 is characterized by: The cutter field depth index FPI = F / P·N, where F represents the cutter head thrust, P represents the cutter head penetration, and N represents the number of cutters on the cutter head. The average friction rate of the positive cutter is Y1 = 0.0025ln X1-0.0929, where Y1 represents the average wear rate of the positive cutter and X1 represents the field depth index FPI.

5. The TBM intelligent decision-making method based on rock quartz content and cutterhead wear according to claim 1 is characterized by: Rock wear resistance index RAI = U × EQC, where U represents the uniaxial compressive strength of rock, EQC represents the equivalent quartz content of rock, and then the rock abrasiveness index CAI is obtained: Y2=0.996ln X2-0.656, Y2 represents the rock abrasiveness index CAI, and X2 represents the rock wear resistance index RAI.

6. The TBM intelligent decision-making method based on rock quartz content and cutterhead wear according to claim 1 is characterized by: The specific process of rock classification according to rock abrasiveness index includes: (1) When CAI < 0.3, the surrounding rock is considered to have no friction. When CAI is between 0.3 and 0.5, the surrounding rock is considered to have low friction. In this case, there is no need to consider the wear of the cutterhead for parameter adjustment. (2) When CAI is between 0.5 and 1.0, the surrounding rock is considered to have medium friction. When CAI is between 1.0 and 2.0, the surrounding rock is considered to have friction. At this time, the operating parameters should be fine-tuned, and the parameter weight considering cutterhead wear is 30%; (3) When CAI is between 2.0 and 4.0, the surrounding rock is considered to have high friction. When CAI is between 4.0 and 6.0, the surrounding rock is considered to have extremely high friction. At this time, the operating parameters should be adjusted, and the parameter weight of the cutterhead wear is considered to be 50%.

7. The TBM intelligent decision-making method based on rock quartz content and cutterhead wear according to claim 1 is characterized by: If the predicted value of the excavation parameter fluctuates by more than the set value within the set time, the system will issue an alarm, and the TBM operator will judge the on-site situation and manually control and take corresponding measures.

8. A TBM intelligent decision-making system based on rock quartz content and cutterhead wear, characterized by: include: The parameter acquisition module is configured to detect rock powder to obtain a spectrum when the TBM is working, analyze the spectrum to obtain rock quartz content data, and obtain TBM cutterhead operation parameters; The data preprocessing module is configured to cache and package the collected data in segments; filter abnormal data to extract the correct parameters of the TBM ascent segment as input layer data; The modeling module is configured to use the support vector machine algorithm model to perform deep learning on the input layer data, so as to establish a rock-machine data feedback model and use the state parameters of the surrounding rock in the stable section as the output layer for prediction; establish a rock wear resistance index based on the state parameters of the surrounding rock in the stable section and the quartz content of the rock, and then obtain the wear degree of the cutter; establish a matrix of the state parameters of the surrounding rock in the stable section and consider the wear degree of the cutter head, and use the support vector machine algorithm model to predict the excavation parameters of the steady-state section; The parameter adjustment module is configured to perform weighted processing on the TBM excavation parameters according to the rock friction classification, and output them to the upper computer visualization interface in real time. At the same time, the system sends instructions to the PLC controller to adjust to the optimal excavation parameters of the TBM stable section; The model update module is configured to periodically update the model according to the defined time and upgrade the training model of the input layer data.

9. A computer-readable storage medium, characterized in that: There are multiple instructions, which are suitable for being loaded by a processor of a terminal device and executing part or all of the steps of a TBM intelligent decision-making method based on rock quartz content and cutterhead wear according to any one of claims 1-7.

10. A terminal device, characterized in that: The invention comprises a processor and a computer-readable storage medium, wherein the processor is used to implement various instructions; and the computer-readable storage medium is used to store multiple instructions, wherein the instructions are suitable for being loaded by the processor and executing part or all of the steps of a TBM intelligent decision-making method based on rock quartz content and cutterhead wear according to any one of claims 1 to 7.

Images