RTBM construction method under complex geological conditions

Abstract

The invention discloses an RTBM construction method under a complex geological condition, belongs to the technical field of tunnel construction, and aims to solve the problems of aggravated cutter wear and low construction efficiency caused by delayed geological information acquisition and inaccurate construction parameter adjustment of a traditional hard rock tunnel boring machine under the complex geological condition. The invention provides a construction method based on multi-source geological data fusion and dynamic optimization. The method comprises the following steps: acquiring geological structure, rock texture and surrounding rock stress data in front of a tunnel face in real time through a multi-source geological detection device, classifying surrounding rock grades and identifying a fracture zone and a high-stress area by utilizing edge calculation nodes and combining a support vector machine algorithm, and dynamically adjusting the thrust, the rotating speed and the propelling speed of a cutterhead; meanwhile, the abrasion loss and the abrasion rate of the tool are monitored, and compensation optimization is conducted on the construction parameters based on a preset threshold value. The method can remarkably improve the safety of complex geological tunnel construction, prolong the service life of the cutter and improve the tunneling efficiency.

Description

Technical Field

[0001] The present invention relates to the technical field of tunnel construction, and more specifically, to a RTBM construction method under complex geological conditions. Background Art

[0002] In the construction of hard rock tunnel boring machines (RTBM), complex geological conditions (such as faults, fracture zones, and high-stress areas) pose severe challenges to construction efficiency and equipment safety. Traditional methods rely on manual experience or a single geological detection method (such as drilling sampling), which has problems such as data acquisition lag and insufficient accuracy, making it difficult to guide the adjustment of construction parameters in a timely manner. For example, the classification of surrounding rock grades is usually based on offline laboratory analysis or local drilling data, which cannot reflect the geological changes in front of the face in real time, resulting in the mismatch between the settings of parameters such as cutterhead thrust and speed and the actual working conditions, which can easily cause abnormal wear of the tool and even machine jam accidents. In addition, the existing technology for monitoring tool wear mostly uses manual inspections or single sensors, which lacks real-time and systematicity, and it is difficult to dynamically compensate for construction parameters. These problems are mainly caused by the following reasons: First, the geological detection method is single, and it is impossible to integrate multi-dimensional data (such as wave velocity, stress, and texture) to comprehensively analyze the geological state; second, data processing relies on centralized servers, and the response speed is insufficient; third, the construction parameter adjustment rules are rigid and are not combined with dynamic wear state optimization. Although some studies have attempted to introduce machine learning algorithms to optimize classification models, the model training sample coverage is insufficient and is not linked to real-time construction parameters, so the actual application effect is limited. Therefore, there is an urgent need for a construction method that can integrate multi-source geological data in real time, dynamically optimize construction parameters, and synchronously compensate for tool wear to cope with the variability and uncertainty of complex geological conditions. Summary of the invention

[0003] An object of the present invention is to provide an RTBM construction method under complex geological conditions, which acquires multi-dimensional geological data in real time and dynamically adjusts construction parameters under complex geological conditions to reduce tool wear and improve construction efficiency.

[0004] In order to achieve these purposes and other advantages of the present invention, according to one aspect of the present invention, the present invention provides a RTBM construction method under complex geological conditions, comprising the following steps: Step 1: Using a multi-source geological detection device to obtain real-time geological structure, rock texture and surrounding rock stress data in front of the tunnel face of the hard rock tunnel boring machine; Step 2: Input the data obtained in step 1 into the edge computing node, use the support vector machine algorithm to classify the surrounding rock grade, and identify the broken zone and high stress area; Step 3: dynamically adjust the construction parameters of the hard rock tunnel boring machine according to the surrounding rock grade classification and the identification results of the broken zone and the high stress area in step 2, wherein the construction parameters include the cutter head thrust, the cutter head rotation speed and the propulsion speed; Step 4: During the construction process, the tool wear amount and tool wear rate are monitored in real time, and wear compensation is performed on the construction parameters according to the tool wear amount and tool wear rate: When the tool wear of any single tool reaches 5~8 mm, reduce the thrust of the cutter head by 10%~20%; When the total tool wear rate is greater than 15%, the advancement speed is dynamically calculated according to the formula V′=V×[1-0.5×(total tool wear rate-0.15)] to make the tool wear rate less than 2 mm / min, where V is the original advancement speed and V′ is the adjusted advancement speed.

[0005] Preferably, the multi-source geological exploration device in step 1 includes: The array ultrasonic detector arranged at the front end of the cutterhead of the hard rock tunnel boring machine obtains the wave velocity distribution data of the rock in front of the tunnel face through phased array scanning technology. The rock stratification, fault location, and crack distribution in front of the tunnel face are obtained by analyzing the wave velocity distribution data to form geological structure data; The distributed optical fiber monitoring system integrated into the front shield of the hard rock tunnel boring machine uses Brillouin optical time-domain reflectometry technology to monitor the strain of the surrounding rock along the way and assist in judging the geological structure based on the distribution and changes of strain data; The multispectral imager installed on the cutterhead support structure of the hard rock tunnel boring machine obtains the hyperspectral image of the tunnel face. By processing and analyzing the image, the rock surface roughness and joint orientation are extracted to form rock texture data; The pressure-displacement composite sensor installed on the propulsion cylinder of the hard rock tunnel boring machine synchronously collects the surrounding rock reaction force and shield deformation data, and calculates the surrounding rock stress data; and The data fusion module receives the geological structure, rock texture and surrounding rock stress data obtained by the ultrasonic detector, the optical fiber monitoring system, the multi-spectral imager and the pressure-displacement composite sensor.

[0006] Preferably, the surrounding rock stress data is calculated by the following steps: The pressure value P measured by the pressure-displacement composite sensor is converted into the cylinder thrust F. The calculation formula is F=P×A, where A is the cross-sectional area of ​​the cylinder piston, m²; Substitute the displacement ΔL into the strain formula ε=ΔL / L 0 , calculate the shield strain ε; L 0 is the initial length of the cylinder, m; The surrounding rock stress σ, MPa, is calculated based on the formula σ=E×ε+η×(dε / dt); where E is the elastic modulus of the surrounding rock, MPa; η is the viscosity coefficient, MPa·s; dε / dt is the strain rate, s⁻¹; By correcting σ with the correction coefficient k, we can obtain σ' = k × σ, where the value range of k is 0.85~1.15; Every 10 minutes, σ' is compared with the strain data measured by the distributed optical fiber monitoring system, and the k value is optimized by the least squares method to ensure that the deviation between the calculated result and the measured data is ≤5%.

[0007] Preferably, the specific steps of surrounding rock grade classification and identification of fracture zones and high stress areas in step 2 are as follows: Data preprocessing: filtering and interpolation of geological structure data, extracting SIFT features after smoothing and enhancing rock texture data, and normalizing surrounding rock stress data with Z-score; Feature selection and fusion: For geological structure data, rock wave velocity, crack ratio, and fault spacing are selected; for rock texture data, roughness and texture directionality are extracted; for surrounding rock stress data, maximum principal stress and stress gradient are selected, and the principal component analysis method is used to reduce the dimension of the selected features and then fuse them into feature vectors; Model training: A large amount of typical sample data of different surrounding rock grades, including fracture zones and high stress areas, was collected and divided into training set and test set in a ratio of 7:3. The radial basis kernel function was selected as the kernel function of the support vector machine, and the support vector machine model was trained using the training set. Classification and recognition: The fused feature vector is input into the trained support vector machine model. The model determines the surrounding rock grade category according to the distance between the feature vector of the input data and different classification hyperplanes, and classifies the surrounding rock grade into IV grade; For the identification of the broken zone, a comprehensive discrimination rule based on the rock wave velocity mutation rate, crack rate threshold and texture feature irregularity is set. When the characteristic value output by the model meets the discrimination rule, the area is judged as a broken zone. For the identification of high stress areas, the maximum principal stress and stress gradient threshold in the surrounding rock stress data are used for judgment. When the maximum principal stress corresponding to the input data exceeds the set high stress threshold and the stress gradient is greater than a certain value, the area is identified as a high stress area.

[0008] Preferably, the construction parameters corresponding to different surrounding rock grades in step 3 are adjusted as follows: Level I surrounding rock is stable surrounding rock, and its corresponding cutterhead thrust is 8000~12000 kN, cutterhead speed is 5~8 r / min, and advancement speed is 60~80 mm / min; Level II surrounding rock is basically stable surrounding rock, and its corresponding cutterhead thrust is 12000~15000 kN, cutterhead speed is 4~6r / min, and advancement speed is 40~60 mm / min; Level III surrounding rock is a rock with poor stability, and its corresponding cutterhead thrust is 15000~18000 kN, the cutterhead speed is 3~5 r / min, and the advancement speed is 20~40 mm / min; Level IV surrounding rock is unstable surrounding rock, and its corresponding cutterhead thrust is 18000~22000 kN, cutterhead speed is 2~4r / min, and advancement speed is 10~20 mm / min; Grade V surrounding rock is extremely unstable surrounding rock, and its corresponding cutterhead thrust is 22000~28000 kN, the cutterhead speed is 1~3r / min, and the advancement speed is 5~10 mm / min.

[0009] Preferably, the construction parameters corresponding to the broken zone area and the high stress area in step 3 are adjusted as follows: In the broken zone, the cutterhead thrust is increased by 10%~30% on the basis of the corresponding surrounding rock grade cutterhead thrust, and the adjusted cutterhead thrust shall not exceed 30000 kN, the cutterhead speed is reduced to 50%~70% of the corresponding surrounding rock grade cutterhead speed, and the advancement speed is reduced to 30%~50% of the advancement speed of the corresponding surrounding rock grade; In high stress areas, the cutterhead thrust increases by 20%~40% based on the cutterhead thrust of the corresponding surrounding rock grade, the cutterhead speed decreases by 10%~20% based on the cutterhead speed of the corresponding surrounding rock grade, and the advancement speed is reduced to 40%~60% of the advancement speed of the corresponding surrounding rock grade.

[0010] Preferably, the method for monitoring the tool wear amount and tool wear rate in step 4 is: A ceramic-encapsulated high-permeability induction coil is embedded inside the cutter of a hard rock tunnel boring machine at a distance of 10 to 15 mm from the cutting edge. A ring-shaped induction detection probe is set at the position of the cutter head corresponding to the cutter. The inner diameter of the induction detection probe is larger than the outer diameter of the cutter. Through 10 to 20 kHz alternating magnetic field excitation, a lock-in amplifier is used to detect the change in electromagnetic induction intensity caused by wear, and the tool wear amount and tool wear rate are calculated through A / D conversion and microprocessor table lookup.

[0011] Preferably, when compensating for wear in step 4, construction parameters need to be dynamically compensated in combination with geological conditions, specifically: when the wear of any single tool reaches 5-8 mm, triggering a reduction in the cutter disc thrust, if it is in a broken zone, the cutter disc thrust is additionally reduced by 5%-10%; when the total tool wear rate is >15%, triggering an adjustment in the advancement speed, if it is in a high stress area, the advancement speed reduction is limited to ≤40%.

[0012] Preferably, when the construction parameter adjustment and wear compensation in the broken zone area are triggered simultaneously, the wear compensation rule is executed preferentially.

[0013] The present invention includes at least the following beneficial effects: The present invention optimizes construction parameters in real time through multi-source data fusion and edge computing, significantly reduces tool wear rate, and improves excavation efficiency; the dynamic adjustment mechanism can avoid equipment jamming caused by geological mutations and improve construction safety. Through the collaborative work of multi-source detection devices, the comprehensiveness of geological data collection is improved, and the accuracy of fault and crack identification is high, providing reliable input for subsequent classification. The correction coefficient optimization is introduced into the surrounding rock stress calculation, and the deviation from the optical fiber monitoring data is controlled within 5%, and the stress distribution analysis is closer to the actual working conditions. Using the support vector machine algorithm combined with multi-dimensional feature classification, the accuracy of surrounding rock grade identification is high, and the misjudgment rate of broken zones is reduced. Through the clear adjustment rules of graded parameters, the matching degree between construction parameters and geological conditions is improved, the tool life is extended, and the fluctuation of advancement speed is reduced. The electromagnetic induction wear monitoring has strong real-time performance, and the compensation rules are linked with the geological state, which reduces the incidence of abnormal wear accidents.

[0014] Other advantages, objectives and features of the present invention will be embodied in part through the following description, and in part will be understood by those skilled in the art through study and practice of the present invention. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] Figure 1 This is a flow chart of the RTBM construction method under complex geological conditions described in one technical solution of the present invention. DETAILED DESCRIPTION

[0016] The present invention is further described in detail below in conjunction with specific implementation modes, so that those skilled in the art can implement the invention with reference to the description.

[0017] It should be understood that the terms such as “having”, “including” and “comprising” used herein do not exclude the existence or addition of one or more other elements or combinations thereof.

[0018] It should be noted that the experimental methods described in the following embodiments are conventional methods unless otherwise specified, and the reagents and materials can be obtained from commercial channels unless otherwise specified.

[0019] like Figure 1 As shown, the present invention provides a RTBM construction method under complex geological conditions, comprising the following steps: Step 1: Using a multi-source geological detection device to obtain real-time geological structure, rock texture and surrounding rock stress data in front of the tunnel face of the hard rock tunnel boring machine; Step 2: Input the data obtained in step 1 into the edge computing node, use the support vector machine algorithm to classify the surrounding rock grade, and identify the broken zone and high stress area; Step 3: dynamically adjust the construction parameters of the hard rock tunnel boring machine according to the surrounding rock grade classification and the identification results of the broken zone and the high stress area in step 2, wherein the construction parameters include the cutter head thrust, the cutter head rotation speed and the propulsion speed; Step 4: During the construction process, the tool wear amount and tool wear rate are monitored in real time, and wear compensation is performed on the construction parameters according to the tool wear amount and tool wear rate: When the tool wear of any single tool reaches 5~8 mm, reduce the thrust of the cutter head by 10%~20%; When the total tool wear rate is greater than 15%, the advancement speed is dynamically calculated according to the formula V′=V×[1-0.5×(total tool wear rate-0.15)] to make the tool wear rate less than 2 mm / min, where V is the original advancement speed and V′ is the adjusted advancement speed.

[0020] In this technical solution, the specific implementation method of the RTBM construction method under complex geological conditions is as follows: In the geological detection link, in order to comprehensively and accurately obtain the geological data in front of the hard rock tunnel boring machine, different types of equipment can be selected for collaborative operation. For example, the Zond-12e geological radar can be used, which has efficient rock structure scanning capabilities and can perform detailed scanning of the range of 10 to 30 meters in front of the face. It can be installed on a specially designed fixed frame at the front end of the cutter shield to ensure that it can directly face the front of the face to carry out scanning work. The XY-2PC advance drilling device can also be used to obtain rock samples. The device can be installed in the middle of the main beam. This position can ensure the stability of the drilling operation and facilitate the acquisition of rock samples from different angles. The TSP-203 acoustic wave detector can be used for surrounding rock stress distribution analysis and installed on the rear supporting frame to reduce the vibration interference of the equipment during the construction process and ensure that the acoustic wave detector can accurately detect the surrounding rock stress. The edge computing node plays a key role in data processing and analysis in the entire construction method, and Advantech UNO-2482G industrial server can be selected. The server has powerful computing power and can carry the TensorFlow framework to implement the support vector machine algorithm. During the data processing process, the system will perform deep learning training on up to 12 characteristic parameters such as quartz content and joint spacing in the surrounding rock sample library. After training with a large amount of sample data, the system can complete the surrounding rock grade classification in a very short time of 200ms. The decision module will then dynamically adjust the construction parameters according to the established algorithms and rules. This ensures that the construction process is compatible with the current geological conditions. The wear of the tool directly affects the construction efficiency and cost, so an accurate tool monitoring system is required. The Keyence LK-G80 laser displacement sensor can be used to measure the tool wear in real time. The sensor is installed in a specific position close to the tool and can accurately measure the wear changes of the tool during the construction process. At the same time, the FLIRA655sc infrared thermometer is used to monitor the tool temperature. The thermometer is installed in a position that can fully cover the tool working area, so as to promptly detect the temperature increase caused by abnormal tool wear. For example, when the wear of a single cutter reaches 6mm, the preset control program in the system will automatically reduce the thrust of the cutter disc by 15% to reduce the working pressure of the cutter. When the total wear rate exceeds 15%, the system will strictly adjust the advancement speed according to the formula V'=V×[1-0.5×(total wear rate-0.15)]. Through the above method, the automated and intelligent construction of hard rock tunnel boring machines can be achieved under complex geological conditions. The multi-source geological detection device obtains geological information in front of the face in real time to provide data support for the adjustment of construction parameters; the dynamic adjustment of construction parameters is optimized according to the surrounding rock grade and the identification results of the broken zone and high stress area to improve construction efficiency and safety; tool wear monitoring and compensation can timely detect tool wear and take compensation measures to extend the service life of the tool.The realization of these technical effects will help reduce construction costs, improve construction quality, shorten construction period, and provide strong technical support for tunnel construction under complex geological conditions.

[0021] In another technical solution, the multi-source geological exploration device in step 1 includes: The array ultrasonic detector arranged at the front end of the cutterhead of the hard rock tunnel boring machine obtains the wave velocity distribution data of the rock in front of the tunnel face through phased array scanning technology. The rock stratification, fault location, and crack distribution in front of the tunnel face are obtained by analyzing the wave velocity distribution data to form geological structure data; The distributed optical fiber monitoring system integrated into the front shield of the hard rock tunnel boring machine uses Brillouin optical time-domain reflectometry technology to monitor the strain of the surrounding rock along the way and assist in judging the geological structure based on the distribution and changes of strain data; The multispectral imager installed on the cutterhead support structure of the hard rock tunnel boring machine obtains the hyperspectral image of the tunnel face. By processing and analyzing the image, the rock surface roughness and joint orientation are extracted to form rock texture data; The pressure-displacement composite sensor installed on the propulsion cylinder of the hard rock tunnel boring machine synchronously collects the surrounding rock reaction force and shield deformation data, and calculates the surrounding rock stress data; and The data fusion module receives the geological structure, rock texture and surrounding rock stress data obtained by the ultrasonic detector, the optical fiber monitoring system, the multi-spectral imager and the pressure-displacement composite sensor.

[0022] In this technical solution, in the selection of geological detection devices, the RITEC RAM-5000 array ultrasonic detector can be used and arranged at the front end of the cutterhead of the hard rock tunnel boring machine. The detector adopts phased array scanning technology, the scanning frequency can be set at 2~5MHz, and the scanning range can cover 5~10 meters in front of the face. By analyzing the rock wave velocity distribution data, when the wave velocity change rate exceeds 10%, it can be judged that there is rock stratification or fault. The Smart distributed optical fiber monitoring system can be used and integrated into the front shield of the hard rock tunnel boring machine. The system adopts Brillouin optical time domain reflection technology, with a monitoring range of one monitoring point every 5 meters, and can monitor the strain of the surrounding rock along the way. When the strain data changes by more than 0.05%, it can assist in judging that the geological structure is abnormal. The multispectral imager can choose the Specim IQ type and be installed on the cutterhead support structure of the hard rock tunnel boring machine. A shock-absorbing bracket can be designed for installing the multispectral imager. The shock-absorbing bracket uses elastic materials (such as rubber, springs, etc.), which can effectively absorb the vibration caused by the rotation of the cutterhead. A buffer pad is set between the bracket and the imager to further reduce the transmission of vibration. A gyroscope can also be used to stabilize the platform to monitor the rotation and vibration of the cutterhead in real time. When vibration is detected, the posture of the imager is adjusted through the motor drive device on the platform to keep a relatively stable shooting angle. The spectral range of the imager is 400~1000 nm, which can obtain hyperspectral images of the tunnel face. The image is processed by image analysis software to extract the surface roughness and joint orientation of the rock. The HBMU9C pressure-displacement composite sensor can be selected and set on the propulsion cylinder of the hard rock tunnel boring machine. The sensor can synchronously collect surrounding rock reaction force and shield deformation data with a measurement accuracy of ±0.1%. The data fusion module can use an embedded fusion module based on FPGA, which receives the geological structure, rock texture and surrounding rock stress data obtained by the above-mentioned devices and performs data fusion processing. The data fusion module can use the weighted average method or the Bayesian network method. The weighted average method is simple and intuitive, and can quickly fuse data; the Bayesian network method can handle the uncertainty and correlation of data, and is suitable for data fusion under complex geological conditions.

[0023] In actual construction, first install the various devices and sensors according to the above positions. During the excavation process, the array ultrasonic detector, distributed fiber optic monitoring system, multi-spectral imager and pressure-displacement composite sensor work continuously, collect data in real time and transmit it to the data fusion module. The data fusion module inputs the processed data into the edge computing node for surrounding rock grade classification and regional identification. Taking a tunnel project as the experimental object, a comparative experimental method was adopted to compare the area using the multi-source geological detection device with the area using the traditional detection method. After statistical analysis, the use of the multi-source geological detection device can improve the accuracy of geological condition judgment, provide a more accurate basis for the dynamic adjustment of subsequent construction parameters, and effectively improve construction efficiency and safety.

[0024] In another technical solution, the surrounding rock stress data is calculated by the following steps: The pressure value P measured by the pressure-displacement composite sensor is converted into the cylinder thrust F. The calculation formula is F=P×A, where A is the cross-sectional area of ​​the cylinder piston, m²; Substitute the displacement ΔL into the strain formula ε=ΔL / L 0 , calculate the shield strain ε; L 0 is the initial length of the cylinder, m; The surrounding rock stress σ, MPa, is calculated based on the formula σ=E×ε+η×(dε / dt); where E is the elastic modulus of the surrounding rock, MPa; η is the viscosity coefficient, MPa·s; dε / dt is the strain rate, s⁻¹; By correcting σ with the correction coefficient k, we can obtain σ' = k × σ, where the value range of k is 0.85~1.15; Every 10 minutes, σ' is compared with the strain data measured by the distributed optical fiber monitoring system, and the k value is optimized by the least squares method to ensure that the deviation between the calculated result and the measured data is ≤5%.

[0025] In this technical solution, in terms of numerical selection, the cross-sectional area A of the cylinder piston is determined according to different models of hard rock tunnel boring machine propulsion cylinders, and the common value range is 0.1~0.5m². The initial length L of the cylinder 0Usually between 1 and 3 m. The elastic modulus E of the surrounding rock can be taken from the common 5~50 GPa according to different geological conditions. Specifically, in the early stage of construction, core samples are taken by drilling on site and sent to the laboratory for uniaxial compression test, triaxial compression test, etc. In the uniaxial compression test, the elastic modulus is determined according to the slope of the initial straight line segment of the stress-strain curve. For samples of different lithologies, tests are carried out separately and a database is established. During the construction process, the corresponding elastic modulus value is selected from the database according to the surrounding rock type classified by the support vector machine model. In-situ testing can also be carried out at the tunnel construction site using equipment such as a borehole elastic modulus instrument. The viscosity coefficient η ranges from 10 to 100 MPa・s. The specific method of obtaining it is to conduct indoor rheological tests on core samples, such as uniaxial compression rheological tests, triaxial compression rheological tests, etc. The viscosity coefficient is determined by measuring the change of strain of the sample under constant load over time and fitting it with a rheological model (such as the Burgers model, etc.). Samples of different lithologies should be tested separately to establish a viscosity coefficient database. Or during the tunnel construction process, displacement monitoring points are arranged to monitor the displacement changes of the surrounding rock in real time. According to the monitoring data, the viscosity coefficient of the surrounding rock is obtained by inversion using the inverse analysis method combined with numerical simulation software (such as FLAC3D, ABAQUS, etc.). This method can dynamically adjust the viscosity coefficient according to the actual situation on site and improve the accuracy of the calculation. In actual construction, the strain rate dε / dt is generally taken in the range of 0.001~0.01s⁻¹. The initial value of the correction coefficient k can be selected within the range of 0.85~1.15. In terms of equipment and component selection, the HBMU9C pressure-displacement composite sensor can be used to measure the pressure value P and the displacement ΔL, which has high measurement accuracy and can meet the construction requirements. The distributed optical fiber monitoring system can select the Smart distributed optical fiber monitoring system to obtain the strain data of the surrounding rock along the way. In terms of computing equipment, an ordinary industrial control computer can be used, equipped with a corresponding data acquisition card and computing software to run the relevant calculation formulas. In terms of material selection, the piston of the propulsion cylinder can be made of high-strength alloy steel to ensure stability and durability under high pressure. In the assembly position, the pressure-displacement composite sensor is installed on the thrust cylinder of the hard rock tunnel boring machine to ensure accurate measurement of the pressure and displacement of the cylinder. The distributed fiber optic monitoring system is integrated into the front shield of the hard rock tunnel boring machine to monitor the strain of the surrounding rock along the way. The working process is that during construction, the pressure-displacement composite sensor measures the pressure value P and displacement ΔL in real time, calculates the cylinder thrust F according to the formula F= P×A, and calculates the cylinder thrust F through ε = ΔL / L 0Calculate the shield strain ε, and then calculate the surrounding rock stress σ according to σ = E×ε+η×(dε / dt), and correct it with the correction coefficient k to get σ'. Every 10 minutes, compare σ' with the strain data measured by the distributed optical fiber monitoring system, and use the least squares method to optimize the k value so that the deviation between the calculated result and the measured data is ≤5%. The use of this method to calculate the surrounding rock stress can effectively improve the calculation accuracy, provide a more reliable basis for the subsequent construction parameter adjustment, and ensure the accurate grasp of the surrounding rock stress changes during the construction process, thereby improving the construction safety and efficiency. By adopting the above-mentioned multi-source geological detection device and surrounding rock stress calculation method, the comprehensiveness and accuracy of geological data collection can be significantly improved. The multi-source geological detection device obtains data from multiple dimensions, making the geological structure, rock texture and surrounding rock stress data more accurate, providing a solid data basis for subsequent construction decisions. In terms of surrounding rock stress calculation, the reliability of the calculation results is improved through rigorous steps and parameter corrections. As a result, the construction team can judge the surrounding rock conditions more accurately and dynamically adjust the construction parameters of the hard rock tunnel boring machine to better adapt parameters such as cutterhead thrust, cutterhead speed and propulsion speed to the actual geological conditions, effectively reduce construction risks, reduce construction delays caused by misjudgment of geological conditions, improve construction efficiency, and ensure the smooth progress of the project.

[0026] In another technical solution, the specific steps of surrounding rock grade classification and identification of fracture zones and high stress areas in step 2 are as follows: Data preprocessing: filtering and interpolation of geological structure data, extracting SIFT features after smoothing and enhancing rock texture data, and normalizing surrounding rock stress data with Z-score; Feature selection and fusion: For geological structure data, rock wave velocity, crack ratio, and fault spacing are selected; for rock texture data, roughness and texture directionality are extracted; for surrounding rock stress data, maximum principal stress and stress gradient are selected, and the principal component analysis method is used to reduce the dimension of the selected features and then fuse them into feature vectors; Model training: A large amount of typical sample data of different surrounding rock grades, including fracture zones and high stress areas, was collected and divided into training set and test set in a ratio of 7:3. The radial basis kernel function was selected as the kernel function of the support vector machine, and the support vector machine model was trained using the training set. Classification and recognition: The fused feature vector is input into the trained support vector machine model. The model determines the surrounding rock grade category according to the distance between the feature vector of the input data and different classification hyperplanes, and classifies the surrounding rock grade into IV grade; For the identification of the broken zone, a comprehensive discrimination rule based on the rock wave velocity mutation rate, crack rate threshold and texture feature irregularity is set. When the characteristic value output by the model meets the discrimination rule, the area is judged as a broken zone. For the identification of high stress areas, the maximum principal stress and stress gradient threshold in the surrounding rock stress data are used for judgment. When the maximum principal stress corresponding to the input data exceeds the set high stress threshold and the stress gradient is greater than a certain value, the area is identified as a high stress area.

[0027] In this technical solution, in the data preprocessing stage, the Butterworth low-pass filter can be used to filter the geological structure data, the cutoff frequency is set to 50Hz, and the missing data is supplemented with the cubic spline interpolation method. The rock texture data is processed by the Gaussian smoothing algorithm, the window size is set to 3×3 pixels, and the SIFT feature extraction algorithm is used to obtain a 128-dimensional feature vector. The surrounding rock stress data adopts the Z-score standardization method, and the calculation formula is (X-μ) / σ, where μ is the sample mean and σ is the sample standard deviation. Data preprocessing can be completed on Advantech UNO-2482G industrial server, which is equipped with Intel Corei7 processor and 16GB memory to ensure data processing efficiency. In the process of feature selection and fusion, the rock wave velocity range is 2000~6000m / s, the fracture rate threshold is set to 0.15~0.45, and the fault spacing is 0.5~5m. The rock roughness is represented by the Ra parameter, with a value range of 0.5~5μm, and the texture directionality is calculated by the gray level co-occurrence matrix to calculate the angle parameter. The surrounding rock stress data selected the maximum principal stress (10~50MPa) and stress gradient (0.1~1.5MPa / m). The principal component analysis method can retain 95% of the cumulative variance contribution rate, and the eigenvector after dimension reduction is input into the edge computing node equipped with the TensorFlow framework for processing. During model training, 500 groups of typical sample data of a tunnel project were collected and divided into training set and test set in a ratio of 7:3. The support vector machine model uses the radial basis kernel function, the parameter γ is set to 0.1, and the penalty coefficient C is set to 10. The surrounding rock grade classification standard refers to the "Railway Tunnel Design Code", the maximum principal stress of Class I surrounding rock is <15MPa, and Class V surrounding rock is >40MPa. The fracture zone discrimination rule is set as: the wave velocity mutation rate is >15% and the crack rate is >0.3, and the texture irregularity exceeds the threshold. The high stress area judgment standard is the maximum principal stress >25MPa and the stress gradient is >0.5MPa / m.

[0028] The above-mentioned surrounding rock grade classification and regional identification methods can significantly improve the analysis accuracy and classification efficiency of geological data. Multi-dimensional data preprocessing technology effectively eliminates noise interference and unifies data formats, providing a reliable basis for subsequent analysis. Feature selection and fusion strategies reduce data redundancy through principal component analysis, retain key geological feature information, and make model input more representative. The support vector machine model combined with typical sample training can accurately distinguish different surrounding rock grades, identify broken zones and high stress areas, and significantly improve the detection rate of abnormal geological areas. The overall solution uses multi-technical collaborative optimization to systematically improve the accuracy of geological classification, regional identification reliability, and construction parameter adaptability, providing a strong guarantee for safe and efficient construction under complex geological conditions.

[0029] In another technical solution, the construction parameters corresponding to different surrounding rock grades in step 3 are adjusted as follows: Level I surrounding rock is stable surrounding rock, and its corresponding cutterhead thrust is 8000~12000 kN, cutterhead speed is 5~8 r / min, and advancement speed is 60~80 mm / min; Level II surrounding rock is basically stable surrounding rock, and its corresponding cutterhead thrust is 12000~15000 kN, cutterhead speed is 4~6r / min, and advancement speed is 40~60 mm / min; Level III surrounding rock is a rock with poor stability, and its corresponding cutterhead thrust is 15000~18000 kN, the cutterhead speed is 3~5 r / min, and the advancement speed is 20~40 mm / min; Level IV surrounding rock is unstable surrounding rock, and its corresponding cutterhead thrust is 18000~22000 kN, cutterhead speed is 2~4r / min, and advancement speed is 10~20 mm / min; Grade V surrounding rock is extremely unstable surrounding rock, and its corresponding cutterhead thrust is 22000~28000 kN, the cutterhead speed is 1~3r / min, and the advancement speed is 5~10 mm / min.

[0030] In this technical solution, the adjustment of construction parameters involves multi-dimensional numerical selection. The thrust of the cutter disc can be set in a gradient range of 8000~28000kN according to the surrounding rock grade, the speed adjustment range is 1~8r / min, and the propulsion speed is controlled at 5~80mm / min. HBM U2A pressure sensor can be used to monitor the thrust of the cutter disc in real time, Omron E6B2-CWZ6C encoder can be used to measure the speed, and Rexroth A4VG variable pump can be used to adjust the propulsion speed. The cutter disc is made of high-strength alloy steel material, with a tungsten carbide wear-resistant layer welded on the surface, and the hydraulic system uses ISO VG46 anti-wear hydraulic oil. Each device is installed in a standard position, the pressure sensor is integrated at the piston rod end of the propulsion cylinder, the encoder is installed at the shaft end of the cutter disc drive motor, and the hydraulic pump group is arranged at the rear supporting frame hydraulic station. When the system identifies the surrounding rock grade, the PLC controller automatically calls the corresponding interval value according to the preset parameter table: stable surrounding rock adopts high-speed fast propulsion mode, and extremely unstable surrounding rock switches to low-speed slow propulsion mode. The parameter adjustment response time is controlled within 300ms to ensure construction continuity. This technical solution achieves fine-grained control of the operating parameters of the tunnel boring machine by establishing a gradient correspondence between the surrounding rock grade and the construction parameters. As the stability of the surrounding rock decreases, the thrust of the cutterhead gradually increases while the rotation speed and propulsion speed decrease accordingly. This dynamic adaptation mechanism significantly improves the equipment's adaptability to complex formations. Practice has shown that this method can effectively reduce the wear rate of the cutter, reduce the construction stagnation time, and improve the tunneling efficiency while ensuring construction safety. It is particularly suitable for engineering applications in weak surrounding rocks and high-stress areas.

[0031] In another technical solution, the construction parameters corresponding to the broken zone area and the high stress area in step 3 are adjusted as follows: In the broken zone, the cutterhead thrust is increased by 10%~30% on the basis of the corresponding surrounding rock grade cutterhead thrust, and the adjusted cutterhead thrust shall not exceed 30000 kN, the cutterhead speed is reduced to 50%~70% of the corresponding surrounding rock grade cutterhead speed, and the advancement speed is reduced to 30%~50% of the advancement speed of the corresponding surrounding rock grade; In high stress areas, the cutterhead thrust increases by 20%~40% based on the cutterhead thrust of the corresponding surrounding rock grade, the cutterhead speed decreases by 10%~20% based on the cutterhead speed of the corresponding surrounding rock grade, and the advancement speed is reduced to 40%~60% of the advancement speed of the corresponding surrounding rock grade.

[0032] In this technical solution, in the parameter adjustment of special geological areas, the thrust of the cutter disc can be appropriately increased on the basis of the corresponding surrounding rock grade in the broken zone area, while the rotation speed and propulsion speed can be significantly reduced. The thrust can be further increased in the high stress area, the rotation speed can be appropriately reduced and the propulsion speed can be slowed down. Rexroth hydraulic system can be used to achieve thrust adjustment, Parker Hannifin proportional valve can be used to control the speed, and OMRON photoelectric sensor can be used to monitor the propulsion speed. The cutter disc is made of high-strength steel with surface strengthening treatment, and the hydraulic system uses fire-resistant hydraulic oil to ensure reliability under extreme working conditions. The equipment installation follows a standardized layout, the pressure sensor is integrated in the propulsion cylinder, the speed encoder is installed in the cutter disc drive motor, and the displacement sensor is arranged at the shield hinge. When the system identifies a special area, the parameter group is automatically switched through the PLC controller, and the auxiliary cooling and support system are started synchronously. The parameter adjustment response time is controlled within a short range to ensure construction continuity. This technical solution significantly improves the adaptability of the tunnel boring machine under complex working conditions by establishing a parameter compensation mechanism for special geological areas. The low-speed and high-thrust mode in the broken zone reduces the disturbance of the surrounding rock, and the parameter combination in the high-stress area reduces the equipment load. Both adjustment strategies can effectively extend the tool life and reduce the construction downtime. Engineering practice shows that this method can improve the construction efficiency in special areas, reduce the frequency of tool replacement, and reduce the construction safety accident rate, providing a more reliable technical guarantee for tunnel engineering.

[0033] In another technical solution, the method for monitoring the tool wear amount and tool wear rate in step 4 is: A ceramic-encapsulated high-permeability induction coil is embedded inside the cutter of a hard rock tunnel boring machine at a distance of 10 to 15 mm from the cutting edge. A ring-shaped induction detection probe is set at the position of the cutter head corresponding to the cutter. The inner diameter of the induction detection probe is larger than the outer diameter of the cutter. Through 10 to 20 kHz alternating magnetic field excitation, a lock-in amplifier is used to detect the change in electromagnetic induction intensity caused by wear, and the tool wear amount and tool wear rate are calculated through A / D conversion and microprocessor table lookup.

[0034] In this technical solution, in the tool wear monitoring system, the induction coil can be wound with TDK PC40 core material, with the number of turns designed to be 50~80 turns, pre-buried in the tool at a distance of 10~15mm from the cutting edge, and encapsulated with 99.9% pure alumina ceramic. The induction detection probe can be a ring-shaped coil produced by Laird, with an inner diameter 2~5mm larger than the outer diameter of the tool, and made of polyetheretherketone (PEEK). The excitation signal uses a 10~20kHz alternating current, which is generated by a RIGOL DG4162 function generator. The phase-locked amplifier can be Stanford Research System SR830, which is used to extract weak electromagnetic signals. The microprocessor can be STM32F407, which has a built-in 12-bit A / D converter and a wear amount lookup algorithm. When the equipment is installed, the induction coil is fixed inside the tool base through a high-temperature brazing process, and the detection probe is installed on the ring bracket of the tool corresponding to the tool of the cutter head through bolts. When the cutter head rotates, the excitation coil generates an alternating magnetic field. The wear of the tool causes the air gap between the coil and the probe to increase, causing the inductance value to change. The phase-locked amplifier amplifies the signal and inputs it into the microprocessor, and calculates the wear amount by table lookup. The wear rate is obtained by dividing the difference between two adjacent measurement values ​​by the excavation distance. The system automatically records data every 50 mm of advancement and stores it in the SD card for subsequent analysis. This technical solution realizes real-time monitoring of tool wear through the principle of electromagnetic induction, and the embedded coil design ensures that the measurement is not affected by harsh environments. The combination of alternating magnetic field excitation and phase-locked amplification technology significantly improves the signal detection sensitivity. This method can obtain wear data without disassembling the tool, providing a reliable basis for adjusting construction parameters, effectively extending the service life of the tool, reducing downtime for tool change, and reducing construction costs.

[0035] In another technical solution, when compensating for wear in step 4, it is necessary to dynamically compensate the construction parameters in combination with the geological conditions, specifically: when the wear of any single tool reaches 5~8 mm, triggering a reduction in the cutter disc thrust, if it is in a broken zone, the cutter disc thrust is additionally reduced by 5%~10%; when the total tool wear rate is >15% and the advancement speed is adjusted, if it is in a high stress area, the advancement speed reduction is limited to ≤40%.

[0036] In this technical solution, the wear compensation system can use Keyence laser displacement sensor to monitor the wear of single cutter, and cooperate with HBM strain pressure sensor to detect the thrust of cutter disc. In the cutter disc structure design, the sensor is installed inside the cutter holder and connected to the cutter disc spindle through an elastic coupling to ensure that the measurement signal is not disturbed by vibration. The control system adopts Siemens S7-1500 PLC, which integrates fuzzy control algorithm module and can receive the surrounding rock status information of geological detection system in real time. When the wear of single cutter reaches the threshold, the system automatically triggers the compensation program. The adjustment of the advancement speed in the high stress area is realized through the speed-stress coupling model to ensure that the necessary excavation efficiency is maintained while reducing the wear of the cutter. This technical solution realizes the fine control of construction parameters by establishing a linkage mechanism between geological state and wear compensation. When the wear of single cutter exceeds the limit and is in the broken zone, the superimposed compensation measures further reduce the equipment load; the high stress area maintains the construction continuity by limiting the reduction of advancement speed. Practice has shown that this method can effectively extend the service life of the cutter, reduce the incidence of sudden failures, and reduce engineering risks while ensuring construction efficiency.

[0037] In another technical solution, when the construction parameter adjustment and wear compensation in the broken zone area are triggered at the same time, the wear compensation rules are executed first. In this technical solution, the dynamic compensation logic is implemented through graded thresholds. When multiple trigger conditions coexist, the system executes the control strategy according to the principle that wear compensation takes precedence over geological adjustment. For example, when the thrust is reduced when the single-blade wear triggers, if the broken zone characteristics (such as sudden changes in acoustic wave velocity or abnormal optical fiber strain) are detected simultaneously, the system will superimpose secondary compensation on the basis of the standard reduction. The priority strategy under multiple trigger conditions avoids parameter conflicts and significantly improves the construction safety of special geological sections.

[0038] Example: A tunnel excavation project is carried out in a mountainous area with complex geological conditions. There are various rock types (such as granite, limestone, sandstone, etc.) in the area, and there may be fracture zones and high stress areas. In order to ensure the efficiency, safety and quality of tunnel construction, the RTBM construction method under complex geological conditions described in the present invention is adopted, as follows: 1. Multi-source geological data collection Ultrasonic detector: A phased array ultrasonic detector is installed at the front end of the shield machine cutterhead to emit ultrasonic signals at a certain time interval (such as every 5 minutes) to detect the geological structure ahead. According to the propagation speed and reflection characteristics of ultrasonic waves in different media, geological stratification information, rock integrity and other data are obtained. For example, when an abnormal rock reflection signal is detected, it may indicate the presence of a broken zone ahead.

[0039] Fiber optic monitoring system: Integrate the distributed fiber optic system into the shield and lining structure of the shield machine. The fiber optic sensor monitors the strain of the surrounding rock in real time by measuring the propagation changes of the optical signal in the optical fiber. For example, during the tunneling process, when the strain value of the optical fiber monitoring in a certain area suddenly increases, it may indicate that the stress of the surrounding rock in that area has changed.

[0040] Multispectral imager: Installed on the cutterhead support structure, it uses its sensitivity to different light spectra to capture images of the rock in front of it. By analyzing the reflection characteristics of the rock under different spectra, it can identify information such as the rock's texture, composition, and degree of weathering. For example, limestone may show a unique reflection pattern under a specific spectrum.

[0041] Pressure-displacement composite sensor: Installed on the thrust cylinder and shield of the shield machine, it measures the pressure of the cylinder and the displacement of the shield. These data are used to calculate the stress and deformation of the surrounding rock. For example, the resistance of the surrounding rock to the shield machine can be inferred based on the change in cylinder pressure.

[0042] 2. Data fusion and processing Data fusion module Weighted average method: According to the reliability and accuracy of each sensor, weights are assigned to the ultrasonic detector, fiber optic monitoring system, multispectral imager and pressure-displacement composite sensor, such as 0.3, 0.2, 0.2, and 0.3. For the data collected at each moment, the weighted average is calculated. For example, for the geological data at a certain moment, the data of each sensor is multiplied by the corresponding weight and added to obtain the fused geological data.

[0043] Bayesian network method: Construct a Bayesian network containing nodes such as geological conditions, tool wear, and tunneling parameters. Determine the conditional probability table between nodes based on historical data and expert knowledge. When new sensor data arrives, use the Bayesian theorem to update the posterior probability of each node to obtain a more accurate fusion result. For example, when the multispectral imager detects a change in rock texture, the judgment of the geological conditions ahead is updated through the Bayesian network.

[0044] Edge computing nodes Data preprocessing: Clean the fused data to remove noise and outliers. For example, use a filtering algorithm to process the data of an ultrasonic detector to remove abnormal signals caused by equipment interference.

[0045] Feature selection and fusion: Extract key features from the preprocessed data, such as rock density, elastic modulus, strain rate, etc., and fuse these features. For example, use the principal component analysis method to reduce the dimensionality of multiple features and extract the comprehensive features that best represent the geological conditions.

[0046] Support vector machine model training: Use historical geological data and corresponding surrounding rock grades as training samples to train the support vector machine model. Adjust the model parameters, such as kernel function type, penalty factor, etc., to improve the classification accuracy of the model. For example, the radial basis kernel function is used for training, and the penalty factor is adjusted after multiple iterations to make the classification accuracy of the model on the validation set reach more than 90%.

[0047] 3. Surrounding rock classification and construction parameter adjustment Surrounding rock grade classification: The processed data is input into the trained support vector machine model to obtain the surrounding rock grade classification results, which are divided into grade I - V surrounding rock. For example, when the model outputs grade III surrounding rock, it means that the stability of the surrounding rock in this area is average.

[0048] Construction parameter adjustment Conventional surrounding rock grade adjustment: According to the surrounding rock grade, adjust the tunneling parameters of the shield machine. For Class I surrounding rock, the propulsion speed and cutterhead speed can be appropriately increased; for Class V surrounding rock, reduce the propulsion speed and cutterhead speed, and increase the propulsion force. For example, in Class I surrounding rock, increase the propulsion speed to 70mm / min and the cutterhead speed to 7r / min; in Class V surrounding rock, reduce the propulsion speed to 50mm / min, reduce the cutterhead speed to 5r / min, and increase the propulsion force to 13000kN.

[0049] Special area adjustments Broken zone area: When the front is identified as a broken zone area, the cutterhead thrust is increased by 20%, while the cutterhead speed is reduced by 60% and the advancement speed is reduced by 40%. For example, if the surrounding rock is level III, the original cutterhead thrust is 16000kN, the cutterhead speed is 4r / min, and the advancement speed is 30mm / min. After entering the broken zone, the cutterhead thrust is increased to 19200kN, the cutterhead speed is reduced to 1.6r / min, and the advancement speed is reduced to 18mm / min.

[0050] High stress area: When a high stress area is detected, the cutterhead thrust is increased by 30%, the cutterhead speed is reduced by 15%, and the advancement speed is reduced by 50%. For example, if the surrounding rock is level III, the original cutterhead thrust is 16000kN, the cutterhead speed is 4r / min, and the advancement speed is 30mm / min. In the high stress area, the cutterhead thrust is increased to 20800kN, the cutterhead speed is adjusted to 3.4r / min, and the advancement speed is adjusted to 15mm / min.

[0051] 4. Tool wear monitoring and compensation Real-time monitoring of tool wear: An electromagnetic induction wear monitoring device is used to monitor tool wear in real time. The device measures tool wear by sensing changes in the induction coil embedded inside the tool. For example, tool wear data is obtained every 10 minutes.

[0052] Wear compensation adjustment Single blade wear compensation: When the single blade wear reaches 5~8mm, reduce the disc thrust by 10%~20%. For example, if the single blade wear reaches 6mm, reduce the disc thrust from 20000kN to 17000kN.

[0053] Total wear rate compensation: Calculate the total wear rate of the tool (wear rate = wear amount / original length of the tool). When the total wear rate exceeds 15%, adjust the feed speed V' = V×(1-0.5Δ), where Δ is the difference between the total wear rate and 15%. For example, when the total wear rate is 20%, if the original feed speed V is 50mm / min, the adjusted feed speed V'=50×(1- 0.5×(0.2-0.15)) = 48.75mm / min.

[0054] 5. Compensation parameter optimization and coordinated adjustment Compensation parameter optimization: According to the surrounding rock stress data and tool wear obtained by the optical fiber monitoring system, the compensation parameters are optimized. For example, by comparing the surrounding rock stress monitored by the optical fiber with the theoretical calculated value, the correction coefficient k is adjusted to optimize the calculation result of the surrounding rock stress.

[0055] 6. Cycle monitoring and adjustment During the entire tunnel construction process, the above steps are repeated continuously, geological conditions and tool wear are monitored in real time, and construction parameters are adjusted in a timely manner based on the monitoring results to form a closed-loop construction control system to ensure the efficiency, safety and quality of tunnel construction.

[0056] Although the embodiments of the present invention have been disclosed as above, they are not limited to the applications listed in the specification and the implementation modes, and they can be fully applied to various fields suitable for the present invention. For those familiar with the art, additional modifications can be easily implemented. Therefore, without departing from the general concept defined by the claims and the scope of equivalents, the present invention is not limited to the specific details and the illustrations shown and described herein.

Claims

1. RTBM construction method under complex geological conditions, characterized by: The following steps are involved: Step 1: Using a multi-source geological detection device to obtain real-time geological structure, rock texture and surrounding rock stress data in front of the tunnel face of the hard rock tunnel boring machine; Step 2: Input the data obtained in step 1 into the edge computing node, use the support vector machine algorithm to classify the surrounding rock grade, and identify the broken zone and high stress area; Step 3: dynamically adjust the construction parameters of the hard rock tunnel boring machine according to the surrounding rock grade classification and the identification results of the broken zone and the high stress area in step 2, wherein the construction parameters include the cutter head thrust, the cutter head rotation speed and the propulsion speed; Step 4: During the construction process, the tool wear amount and tool wear rate are monitored in real time, and wear compensation is performed on the construction parameters according to the tool wear amount and tool wear rate: When the tool wear of any single cutter reaches 5~8 mm, reduce the thrust of the cutter head by 10%~20%; When the total tool wear rate is greater than 15%, the advancement speed is dynamically calculated according to the formula V′=V×[1-0.5×(total tool wear rate-0.15)] to make the tool wear rate less than 2 mm / min, where V is the original advancement speed and V′ is the adjusted advancement speed.

2. The RTBM construction method under complex geological conditions as claimed in claim 1, characterized in that: The multi-source geological exploration device in step 1 includes: The array ultrasonic detector arranged at the front end of the cutterhead of the hard rock tunnel boring machine obtains the wave velocity distribution data of the rock in front of the tunnel face through phased array scanning technology. The rock stratification, fault location, and crack distribution in front of the tunnel face are obtained by analyzing the wave velocity distribution data to form geological structure data; The distributed optical fiber monitoring system integrated into the front shield of the hard rock tunnel boring machine uses Brillouin optical time-domain reflectometry technology to monitor the strain of the surrounding rock along the way and assist in judging the geological structure based on the distribution and changes of strain data; The multispectral imager installed on the cutterhead support structure of the hard rock tunnel boring machine obtains the hyperspectral image of the tunnel face. By processing and analyzing the image, the rock surface roughness and joint orientation are extracted to form rock texture data; The pressure-displacement composite sensor installed on the propulsion cylinder of the hard rock tunnel boring machine synchronously collects the surrounding rock reaction force and shield deformation data, and calculates the surrounding rock stress data; and The data fusion module receives the geological structure, rock texture and surrounding rock stress data obtained by the ultrasonic detector, the optical fiber monitoring system, the multi-spectral imager and the pressure-displacement composite sensor.

3. The RTBM construction method under complex geological conditions as claimed in claim 2, characterized in that: The surrounding rock stress data is calculated by the following steps: The pressure value P measured by the pressure-displacement composite sensor is converted into the cylinder thrust F. The calculation formula is F=P×A, where A is the cross-sectional area of ​​the cylinder piston, m²; Substitute the displacement ΔL into the strain formula ε=ΔL / L0 to calculate the shield strain ε; L0 is the initial length of the cylinder, m; The surrounding rock stress σ, MPa, is calculated based on the formula σ=E×ε+η×(dε / dt); where E is the elastic modulus of the surrounding rock, MPa; η is the viscosity coefficient, MPa·s; dε / dt is the strain rate, s⁻¹; By correcting σ with the correction coefficient k, we can obtain σ' = k × σ, where the value range of k is 0.85~1.15; Every 10 minutes, σ' is compared with the strain data measured by the distributed optical fiber monitoring system, and the k value is optimized by the least squares method to ensure that the deviation between the calculated result and the measured data is ≤5%.

4. The RTBM construction method under complex geological conditions as claimed in claim 1, characterized in that: The specific steps of surrounding rock grade classification and identification of fracture zones and high stress areas in step 2 are as follows: Data preprocessing: Filter and interpolate geological structure data, extract SIFT features after smoothing and enhancing rock texture data, and standardize surrounding rock stress data with Z-score; Feature selection and fusion: For geological structure data, rock wave velocity, fracture ratio, and fault spacing are selected; for rock texture data, roughness and texture directionality are extracted; The maximum principal stress and stress gradient are selected from the surrounding rock stress data, and the selected features are reduced in dimension by principal component analysis and then fused into feature vectors; Model training: A large amount of typical sample data of different surrounding rock grades, including fracture zones and high stress areas, was collected and divided into training set and test set in a ratio of 7:

3. The radial basis kernel function was selected as the kernel function of the support vector machine, and the support vector machine model was trained using the training set. Classification and recognition: The fused feature vector is input into the trained support vector machine model. The model determines the surrounding rock grade category according to the distance between the feature vector of the input data and different classification hyperplanes, and classifies the surrounding rock grade into IV grade; For the identification of the broken zone, a comprehensive discrimination rule based on the rock wave velocity mutation rate, crack rate threshold and texture feature irregularity is set. When the characteristic value output by the model meets the discrimination rule, the area is judged as a broken zone. For the identification of high stress areas, the maximum principal stress and stress gradient threshold in the surrounding rock stress data are used for judgment. When the maximum principal stress corresponding to the input data exceeds the set high stress threshold and the stress gradient is greater than a certain value, the area is identified as a high stress area.

5. The RTBM construction method under complex geological conditions as claimed in claim 1, characterized in that: The construction parameters corresponding to different surrounding rock grades in step 3 are adjusted as follows: Level I surrounding rock is stable surrounding rock, and its corresponding cutterhead thrust is 8000~12000 kN, cutterhead speed is 5~8 r / min, and advancement speed is 60~80 mm / min; Level II surrounding rock is basically stable surrounding rock, and its corresponding cutterhead thrust is 12000~15000 kN, cutterhead speed is 4~6 r / min, and advancement speed is 40~60 mm / min; Level III surrounding rock is a rock with poor stability, and its corresponding cutterhead thrust is 15000~18000 kN, the cutterhead speed is 3~5r / min, and the advancement speed is 20~40 mm / min; Level IV surrounding rock is unstable surrounding rock, and its corresponding cutterhead thrust is 18000~22000 kN, cutterhead speed is 2~4 r / min, and advancement speed is 10~20 mm / min; Grade V surrounding rock is extremely unstable surrounding rock, and its corresponding cutterhead thrust is 22000~28000 kN, the cutterhead speed is 1~3 r / min, and the advancement speed is 5~10 mm / min.

6. The RTBM construction method under complex geological conditions as claimed in claim 1, characterized in that: The construction parameters corresponding to the broken zone area and the high stress area in step 3 are adjusted as follows: In the broken zone, the cutterhead thrust is increased by 10%~30% on the basis of the corresponding surrounding rock grade cutterhead thrust, and the adjusted cutterhead thrust shall not exceed 30000 kN, the cutterhead speed is reduced to 50%~70% of the corresponding surrounding rock grade cutterhead speed, and the advancement speed is reduced to 30%~50% of the advancement speed of the corresponding surrounding rock grade; In high stress areas, the cutterhead thrust increases by 20%~40% based on the cutterhead thrust of the corresponding surrounding rock grade, the cutterhead speed decreases by 10%~20% based on the cutterhead speed of the corresponding surrounding rock grade, and the advancement speed is reduced to 40%~60% of the advancement speed of the corresponding surrounding rock grade.

7. The RTBM construction method under complex geological conditions as claimed in claim 6, characterized in that: The monitoring method of tool wear amount and tool wear rate in step 4 is: A ceramic-encapsulated high-permeability induction coil is embedded inside the cutter of a hard rock tunnel boring machine at a distance of 10 to 15 mm from the cutting edge. A ring-shaped induction detection probe is set at the position of the cutter head corresponding to the cutter. The inner diameter of the induction detection probe is larger than the outer diameter of the cutter. Through 10 to 20 kHz alternating magnetic field excitation, a lock-in amplifier is used to detect the change in electromagnetic induction intensity caused by wear, and the tool wear amount and tool wear rate are calculated through A / D conversion and microprocessor table lookup.

8. The RTBM construction method under complex geological conditions as claimed in claim 7, characterized in that: When compensating for wear in step 4, it is necessary to dynamically compensate the construction parameters in combination with the geological conditions. Specifically, when the wear of any single tool reaches 5-8 mm, the cutter disc thrust is reduced. If it is in the broken zone, the cutter disc thrust is further reduced by 5%-10%; when the total tool wear rate is > 15%, the advancement speed is adjusted. If it is in the high stress area, the advancement speed reduction is limited to ≤40%.

9. The RTBM construction method under complex geological conditions as claimed in claim 8, characterized in that: When construction parameter adjustment and wear compensation in the broken zone area are triggered simultaneously, the wear compensation rules are executed first.

Images