Hard rock tunneling machine cutter integration state monitoring, early warning and damping system and method

Abstract

This invention relates to the field of hard rock tunneling equipment monitoring technology, and discloses an integrated status monitoring, early warning, and vibration reduction system and method for hard rock tunneling machine cutters. The system includes a vibration signal acquisition module, a magnetorheological vibration reduction monitoring module, and a data processing and early warning module. The vibration signal acquisition module captures the mechanical response of the cutter in real time; the magnetorheological vibration reduction monitoring module uses a magnetorheological damper to perform active vibration reduction and synchronously acquires the excitation current signal under load through a current monitoring unit. The data processing and early warning module receives signals via a CAN bus, uses an interpolation algorithm to achieve multi-source data time synchronization, fuses mechanical and electromagnetic domain features, and inputs them into a support vector machine model to determine the wear state. The system also dynamically corrects the alarm threshold based on rock hardness and executes graded early warning or shutdown commands based on the judgment results. This invention achieves the integration of vibration reduction and monitoring functions, improving the accuracy and robustness of cutter status identification in complex geological environments.

Description

Technical Field

[0001] This invention relates to the field of hard rock tunneling equipment monitoring technology, specifically to an integrated status monitoring, early warning and vibration reduction system and method for hard rock tunneling machine cutters. Background Technology

[0002] Hard rock tunnel boring machines (TBMs) are core equipment for tunnel engineering and underground space development. Their cutting tool system comes into direct contact with the rock and operates under harsh conditions of high load and strong impact for extended periods, making them highly susceptible to severe wear or fracture. The condition of the cutting tools directly affects tunneling efficiency and construction safety; therefore, developing efficient monitoring and protection technologies is a key focus for the industry.

[0003] In existing hard rock tunnel boring machine (TBM) technology systems, vibration damping and monitoring functions are typically treated as independent subsystems. Vibration damping devices primarily passively or semi-actively attenuate mechanical impacts, while condition monitoring relies on additional, independently installed sensor networks. This decentralized architecture not only increases the complexity of the system hardware and maintenance costs but also leads to data fragmentation. This results in the monitoring system failing to effectively utilize the load feedback information generated by the vibration damping components during operation, leading to low system integration and difficulty in achieving collaborative operation.

[0004] In terms of monitoring methods, existing technologies mainly rely on vibration signal analysis or intermittent visual inspection. However, the operating environment of hard rock tunnel boring machines is extremely complex. The variability of geological conditions leads to strong background noise, and simple vibration signals often mix with various components such as rock breaking noise, mechanical transmission vibration, and stratum feedback, which can easily mask early, subtle wear characteristics. At the same time, traditional discrimination logic often uses a fixed threshold strategy, which is difficult to adapt to working conditions with drastic changes in rock hardness. When the tunnel boring machine enters high-hardness strata, the foundation vibration level naturally increases, and fixed thresholds are prone to triggering false alarms, misjudging normal high-load operation as tool failure; conversely, in soft rock strata, it may cause missed alarms, resulting in limited recognition accuracy.

[0005] Furthermore, due to the lack of an effective real-time online early warning mechanism, on-site maintenance often relies on periodic manual inspections. This lagging maintenance method cannot capture the abrupt changes in tool wear in real time, frequently leading to unplanned downtime and severely impacting construction progress. When attempting to introduce multi-source sensors for comprehensive monitoring, existing technologies often lack precise time synchronization mechanisms for signals of different frequencies (such as high-frequency vibration and low-frequency current), resulting in data discrepancies in timing. In addition, the high dust and humidity environment underground poses stringent challenges to the sealing performance of precision equipment, making it difficult for conventional designs to guarantee the stability and signal-to-noise ratio of long-term monitoring data. Summary of the Invention

[0006] To address the shortcomings of existing technologies, this invention provides an integrated status monitoring, early warning, and vibration reduction system and method for hard rock tunnel boring machine cutters. It aims to solve the problems in existing technologies, such as low system integration due to the separation of vibration reduction and monitoring functions, low identification accuracy and high false alarm rate due to the inability of single signal monitoring combined with fixed threshold discrimination to adapt to complex geological conditions, and the lack of multi-source data synchronization mechanism and high-level protective structure affecting long-term monitoring stability.

[0007] To achieve the above objectives, the present invention provides an integrated status monitoring, early warning and vibration reduction system for hard rock tunneling machine cutters. The system includes a vibration signal acquisition module, a magnetorheological vibration reduction monitoring module and a data processing and early warning module.

[0008] The vibration signal acquisition module is located at key mechanical nodes of the hard rock tunneling machine to capture the mechanical response generated by the cutting tool operation in real time and output multi-dimensional vibration signals. The magnetorheological vibration damping monitoring module is installed on the vibration transmission path and consists of a magnetorheological vibration damper and a current monitoring unit. The magnetorheological vibration damper adjusts the excitation current in response to external control commands to output variable damping force, and the current monitoring unit synchronously captures the output excitation current signal generated by the excitation coil when the magnetorheological vibration damper is subjected to mechanical load. The data processing and early warning module establishes a time-synchronized communication connection with the vibration signal acquisition module and the magnetorheological vibration damping monitoring module through a CAN bus. This module performs feature fusion of the received multi-dimensional vibration signals and the output excitation current signals to determine the tool wear status, and sends graded early warning commands to external alarm devices or shutdown commands to the main control system of the hard rock tunneling machine based on the determination results.

[0009] Preferably, the vibration signal acquisition module includes multiple piezoelectric sensors distributed at different mechanical nodes of the hard rock tunnel boring machine: a vibration monitoring sensor one rigidly installed at the front end of the cutter mounting seat of the hard rock tunnel boring machine's cutting head to acquire the three-dimensional spatial vibration signal of the cutter; a vibration monitoring sensor two installed at the bearing seat of the main shaft of the hard rock tunnel boring machine to acquire the radial vibration data of the main shaft; and a vibration monitoring sensor three installed on the inner wall of the transmission box housing of the hard rock tunnel boring machine to acquire far-field vibration signals.

[0010] In one specific embodiment, the magnetorheological damper adopts a three-stage sealing structure: from the inside to the outside, a perfluoroelastomer directional lip seal and a polytetrafluoroethylene-coated carbon fiber V-ring are sequentially arranged at the port of the magnetorheological damper cylinder; a wave spring, in conjunction with the perfluoroelastomer directional lip seal, provides mechanical preload; an annular buffer cavity is provided between the perfluoroelastomer directional lip seal and the polytetrafluoroethylene-coated carbon fiber V-ring, and the cavity is filled with silicone-based grease containing nano-ceramic particles; the damper shaft passes through this sealing assembly and reciprocates within the magnetorheological damper cylinder.

[0011] The data processing and early warning module includes a data acquisition system and an audible and visual alarm. The data acquisition system has a built-in signal conditioning circuit and an industrial control computer, which synchronously receives vibration signals collected by various vibration monitoring sensors and current signals collected by current monitoring units via a CAN bus. The audible and visual alarm is connected to the data acquisition system and issues a tiered alarm when it receives an abnormal command.

[0012] The data acquisition system adopts the following data synchronization method: the master node broadcasts a global synchronization clock frame; the vibration signal acquisition module and the current monitoring unit act as slave nodes, calibrate their local clocks after receiving the synchronization frame, and embed a timestamp based on a unified time base in the header of each uploaded data packet; the data acquisition system parses the timestamp, uses the time axis of the multidimensional vibration signal as a reference, and uses an interpolation algorithm to perform time compensation on the output excitation current signal, mapping the output excitation current signal to the time point of the multidimensional vibration signal.

[0013] Furthermore, the features extracted by the data processing and early warning module include mechanical domain features and electromagnetic domain features. Mechanical domain features include the maximum absolute value of the signal within a preset time window (vibration peak value), the root mean square value of the signal, and the ratio of the vibration peak value to the root mean square value (peak factor). Electromagnetic domain features include the ratio of the standard deviation to the arithmetic mean of the signal within a preset time window (current amplitude fluctuation coefficient), and the difference between the actual main frequency of the signal and the reference excitation frequency (frequency deviation value).

[0014] In the state discrimination logic, the data processing and early warning module concatenates and normalizes the mechanical domain features and electromagnetic domain features to construct a joint feature vector. This vector is then input into a pre-set tool state discrimination model. This model is based on the support vector machine algorithm and uses the Gaussian radial basis kernel function to map the joint feature vector to a high-dimensional space. By solving the geometric interval between the sample points and the decision boundary, it outputs the tool wear state category, including new tool, light wear, moderate wear, and heavy wear.

[0015] Preferably, the data processing and early warning module has a dynamic threshold calibration function that adapts to working conditions: it reads the current rock hardness parameters of the hard rock tunneling machine and corrects the basic alarm threshold using a mapping model. The correction rule is: based on the currently monitored rock hardness, as the rock hardness increases, the vibration peak threshold is increased proportionally, and the current amplitude fluctuation coefficient threshold is increased proportionally.

[0016] The system executes the following hierarchical early warning logic: combining the tool wear status category with the corrected basic alarm threshold; when the model determines that the wear is moderate and the real-time signal characteristic parameters do not exceed the corrected basic alarm threshold, it is determined to be a level one alarm condition, and the yellow indicator light of the audible and visual alarm flashes and emits an intermittent alarm sound; when the model determines that the wear is severe, or the real-time signal characteristic parameters exceed the corrected basic alarm threshold, it is determined to be a level two alarm condition, and the red indicator light of the audible and visual alarm stays on and emits a continuous alarm sound, while simultaneously outputting a stop signal to the main control PLC of the hard rock tunneling machine.

[0017] A second aspect of this invention provides an integrated method for monitoring, early warning, and vibration reduction of the cutting tools of a hard rock tunnel boring machine, employing the aforementioned system and including the following steps:

[0018] After the system is started with the hard rock tunneling machine, the data processing and early warning module loads the preset tool state discrimination model and threshold parameters, the magnetorheological vibration reduction monitoring module enters the working state, and each sensor executes the self-test program;

[0019] During the tunneling operation, the vibration signal acquisition module collects the multi-dimensional vibration signal of the cutting tool in real time, the magnetorheological damper attenuates the tunneling vibration, and the current monitoring unit synchronously collects the output excitation current signal of the magnetorheological damper when it is working.

[0020] After the collected multidimensional vibration signals are filtered and amplified, they are synchronously transmitted to the data processing and early warning module via the CAN bus along with the output excitation current signal. The signal is then processed to remove power frequency interference and pulse noise.

[0021] The data processing and early warning module extracts the peak value, root mean square value, peak factor of the multidimensional vibration signal, and amplitude fluctuation coefficient and frequency deviation value of the output excitation current signal. The extracted feature parameters are input into the tool condition discrimination model for calculation to determine the current wear state of the tool.

[0022] Based on the judgment results, a graded early warning strategy is implemented: if the signal does not exceed the dynamic threshold but the model determines that it is moderately worn, a level one audible and visual alarm is issued; if the signal exceeds the dynamic threshold or the model determines that it is severely worn, a level two audible and visual alarm is issued and a shutdown command is given.

[0023] This invention provides an integrated condition monitoring, early warning, and vibration reduction system and method for hard rock tunnel boring machine cutters, which has the following beneficial effects:

[0024] 1. This invention changes the traditional single-mode monitoring that relies solely on vibration signals. It utilizes a magnetorheological vibration damping monitoring module that combines vibration damping execution and load feedback functions. It deeply integrates the direct mechanical response (vibration peak value, root mean square) collected by the vibration monitoring sensor with the electromagnetic response (excitation current amplitude fluctuation, frequency deviation) of the magnetorheological damper under variable load conditions. It uses a support vector machine algorithm to process multidimensional heterogeneous data and cross-validates the vibration signal through the fluctuation characteristics of the current signal. This effectively eliminates random vibration interference caused solely by geological inhomogeneity, thereby significantly improving the accuracy of identifying tool wear status, especially early minor wear, in the complex environment of hard rock tunneling.

[0025] 2. The data processing and early warning module of this invention can read rock hardness parameters in real time and automatically correct alarm thresholds using a mapping model. When tunneling through high-hardness rock strata causes the overall foundation vibration level of the system to rise naturally, the system proportionally increases the discrimination thresholds for vibration peak value and current fluctuation coefficient. This effectively distinguishes the signal differences between normal high-load tunneling and abnormal tool wear, avoids the frequent false alarms caused by fixed threshold strategies when operating in variable-hardness strata, and reduces unnecessary shutdown inspections while ensuring equipment safety.

[0026] 3. This invention solves the problem of asynchronous sampling between high-frequency vibration signals and low-frequency current signals by using the master node broadcasting and interpolation algorithm of the data acquisition system. This ensures that the time base of the input data of the feature fusion algorithm is strictly aligned, avoiding discrimination errors caused by timing misalignment. At the same time, for the high dust environment downhole, the magnetorheological vibration damper adopts a three-stage sealing structure of perfluoroelastomer directional lip seal ring, polytetrafluoroethylene V-ring, and grease containing nano-ceramic particles. While blocking the intrusion of fine dust and extending the service life of the vibration damper, it ensures that the excitation current signal, which is the key feedback source, maintains a high signal-to-noise ratio for a long time, thus ensuring the long-term stability of the monitoring system. Attached Figure Description

[0027] Figure 1 This is a schematic diagram of the method flow of the present invention;

[0028] Figure 2 This is a system architecture diagram of the present invention;

[0029] Figure 3 This is a schematic diagram of the magnetorheological fluid cavity sealing device of the present invention;

[0030] Figure 4 This is a schematic diagram of the annular buffer cavity structure of the present invention.

[0031] The components include: 1. Magnetorheological damping cylinder; 2. Perfluoroelastomer directional lip seal; 3. Wave spring; 4. PTFE-coated carbon fiber V-ring; 5. Annular buffer cavity; 6. Damper shaft; 7. Tunneling head; 8. Vibration monitoring sensor one; 9. Magnetorheological damper; 10. Data acquisition system; 11. Current monitoring unit; 12. Audible and visual alarm; 13. Vibration monitoring sensor two; 14. Main shaft; 15. Vibration monitoring sensor three; 100. Vibration signal acquisition module; 200. Magnetorheological damping monitoring module; 300. Data processing and early warning module. Detailed Implementation

[0032] The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0033] See attached document Figure 2 -Appendix Figure 4 This invention provides an integrated status monitoring, early warning, and vibration reduction system for hard rock tunnel boring machine cutters based on magnetorheological technology. The system mainly includes a vibration signal acquisition module 100, a magnetorheological vibration reduction monitoring module 200, and a data processing and early warning module 300. The modules interact and transmit commands via a CAN bus, forming a closed-loop control network.

[0034] The vibration signal acquisition module 100 consists of multiple piezoelectric sensors distributed at key mechanical nodes of the hard rock tunneling machine. The vibration monitoring sensor 8 is rigidly mounted on the front end of the tool mounting base near the tunneling head 7, and is used to collect direct impact vibration signals generated during tool cutting operations.

[0035] The vibration signal acquisition module 100 consists of multiple piezoelectric vibration sensors distributed at key mechanical nodes of the hard rock tunneling machine, used to capture mechanical response signals during cutter operation from all directions. Vibration monitoring sensor 18 is rigidly mounted at the front end of the cutter mounting seat on the tunneling head 7 of the hard rock tunneling machine, at this position closest to the cutting point, and is used to directly acquire the impact vibration signals generated when the cutter cuts the rock. Vibration monitoring sensor 213 is located at the bearing seat of the spindle 14, used to acquire the radial runout and axial movement characteristics of the spindle 14 during rotation, reflecting the stability of the spindle 14. Vibration monitoring sensor 315 is installed on the inner wall of the transmission housing, used to monitor the transmission and attenuation of vibration signals in the mechanical transmission link. All of the above sensors are connected to cables via waterproof sealed joints to adapt to the high humidity and high dust working environment underground.

[0036] The magnetorheological vibration damping monitoring module 200 is the core execution and sensing unit of the system, which includes a magnetorheological vibration damper 9 and a current monitoring unit 11. The magnetorheological vibration damper 9 is installed at the connection between the tunneling head 7 of the tunneling machine and the rear support structure, serving as a key damping element in the vibration transmission path. The current monitoring unit 11 uses a high-precision Hall current sensor, whose circuit is connected to the electromagnetic magnetic field generator inside the magnetorheological vibration damper 9, for real-time acquisition of the output current signal of the excitation coil.

[0037] To ensure long-term reliability under hard rock conditions, the magnetorheological damper 9 employs a dedicated three-stage sealing structure. At the port of the magnetorheological damper cylinder 1, a perfluoroelastomer directional lip seal 2 and a polytetrafluoroethylene (PTFE)-coated carbon fiber V-ring 4 are sequentially arranged from the inside out. The perfluoroelastomer directional lip seal 2 works in conjunction with a wave spring 3, whose mechanical preload ensures a tight fit of the main sealing lip under high temperature and reciprocating motion. An annular buffer cavity 5 is constructed between the perfluoroelastomer directional lip seal 2 and the PTFE-coated carbon fiber V-ring 4. This annular buffer cavity 5 is filled with a silicone-based grease containing nano-ceramic particles, which serves both to lubricate the damper shaft 6 and to absorb any trace dust particles that may intrude. The damper shaft 6 passes through the aforementioned sealing assembly and reciprocates axially within the magnetorheological damper cylinder 1, generating damping force by shearing the carbonyl iron powder-based magnetorheological fluid filling the cavity.

[0038] The data processing and early warning module 300 includes a data acquisition system 10 and an audible and visual alarm 12. The data acquisition system 10 has a built-in signal conditioning circuit and an industrial control computer, responsible for synchronously acquiring and processing the vibration signals transmitted by the vibration signal acquisition module 100 and the current signals transmitted by the magnetorheological vibration reduction monitoring module 200. The data acquisition system 10 ensures the time synchronization of multiple sensor signals via the CAN bus protocol and, based on the processing results, sends excitation current adjustment commands to the magnetorheological vibration reduction monitoring module 200 or early warning commands to the audible and visual alarm 12. The audible and visual alarm 12 is connected to the data acquisition system 10 and is used to issue graded alarm signals upon receiving abnormal commands.

[0039] See attached document Figure 1 -Appendix Figure 4 This invention provides an integrated method for monitoring, early warning, and vibration reduction of cutting tools in hard rock tunnel boring machines. This method is implemented based on the aforementioned hardware system and specifically includes the following steps:

[0040] S1. System initialization: After the system starts with the hard rock tunneling machine, the data processing and early warning module 300 loads the preset tool state discrimination model and threshold parameters, the magnetorheological vibration reduction monitoring module 200 enters the pre-vibration reduction mode, and each sensor executes the self-test program.

[0041] S2. Synchronous signal acquisition: During the tunneling operation, the vibration signal acquisition module 100 acquires the multi-dimensional vibration signal of the cutting tool in real time, and the magnetorheological vibration reduction monitoring module 200 adjusts the magnetic field strength according to the vibration feedback to maintain the active vibration reduction state. At the same time, the current monitoring unit 11 synchronously acquires the output excitation current signal of the magnetic field generator.

[0042] S3. Data transmission and preprocessing: After the collected vibration signal is filtered and amplified, it is transmitted synchronously with the current signal to the data processing and early warning module 300 via the CAN bus, and the signal is processed to remove power frequency interference and pulse noise.

[0043] S4. Feature Extraction and State Judgment: The data processing and early warning module 300 extracts the peak value, root mean square, peak factor of the vibration signal and the amplitude fluctuation coefficient and frequency deviation value of the current signal, respectively. The extracted feature parameters are input into the tool state judgment model for calculation to determine the current wear state of the tool.

[0044] S5. Early warning and feedback control: The control strategy is executed according to the judgment result. If the judgment signal exceeds the preset threshold, the early warning execution unit issues an audible and visual alarm and issues a stop command. If the judgment signal does not exceed the preset threshold, the system maintains a continuous monitoring state, and the magnetorheological vibration reduction monitoring module 200 adaptively outputs control signals to maintain the vibration reduction operation.

[0045] The above steps will be described in detail below with reference to specific embodiments and accompanying drawings.

[0046] In step S1, the system executes the initialization program synchronously with the start of the hard rock tunnel boring machine, which specifically includes the following sub-steps:

[0047] S101. Model Parameter Loading and Data Flow Configuration. After power-on, the data processing and early warning module 300 reads the configuration file in memory. The system retrieves the pre-trained tool condition discrimination model, which contains support vectors and decision function parameters used to distinguish between normal, lightly worn, moderately worn, and heavily worn tool conditions. Simultaneously, the system loads a basic alarm threshold set, which strictly includes the vibration signal amplitude threshold, frequency threshold, and magnetorheological output current fluctuation threshold specified in the handover document. These parameters serve as the comparison benchmark for subsequent condition discrimination, ensuring that the system has clear judgment criteria from the outset of operation.

[0048] S102. Electrical self-test of the sensor array. Simultaneously with parameter loading, the data acquisition system sends a status query command to the front-end sensor network. For the piezoelectric vibration sensors distributed at the tool mount, spindle 14 bearing, and transmission box, the system checks whether their static bias voltage is within the linear operating range to rule out open-circuit or short-circuit faults. For the Hall current sensor in the magnetorheological vibration damping monitoring module 200, the system checks whether its zero-point drift value is within the calibration range. If any sensor reports an abnormality, the data processing and early warning module 300 will block subsequent processes and report a fault code; if the entire system self-test passes, the system is marked to enter standby monitoring mode. The specific circuit detection principles of the above sensors are well-known to those skilled in the art and will not be elaborated upon here.

[0049] S103. Construction of the Magnetorheological Pre-damping Mode. After receiving the ready command, the magnetorheological vibration monitoring module 200 drives the electromagnetic magnetic field generator to output an initial excitation current of a preset intensity. With the current flowing into the excitation coil, a steady-state magnetic field is established in the magnetorheological fluid chamber inside the magnetorheological damper 9. Under the action of this magnetic field, the carbonyl iron powder magnetic particles in the magnetorheological fluid align along the magnetic field lines, forming a chain-like structure. Through this change in microstructure, the rheological properties of the magnetorheological fluid change from an initial Newtonian fluid state to a Bingham fluid state with shear yield stress. This process enables the magnetorheological damper 9 to possess basic damping force before being subjected to external mechanical impact, thereby absorbing overload impact energy the moment the cutter cuts into the rock wall and preventing cutter chipping.

[0050] After system initialization and the establishment of pre-damping status are completed in step S1, the system proceeds to the signal synchronization acquisition stage in step S2. During this stage, the system executes the following sub-steps:

[0051] S201. Three-dimensional vibration signal acquisition at the cutting source. Vibration monitoring sensor 8, located at the front end of the tool mounting base of the tunneling head 7, serves as the main acquisition source, acquiring the three-dimensional spatial vibration signal of the tool in real time at a sampling frequency of 500Hz. According to the definition in this embodiment: the X-axis (axial direction) is parallel to the center line of the tool spindle 14 and points in the cutting direction. The signal amplitude in this direction reflects the axial impact force and edge wear state of the tool; the Y-axis (radial direction) is perpendicular to the center line of the spindle 14 and points radially outward. The signal in this direction reflects the radial load imbalance and installation deviation; the Z-axis (circumferential direction) is along the tangent direction of tool rotation. The signal in this direction reflects the fluctuation of circumferential cutting resistance and rotational stability.

[0052] S202. Spindle 14 rotational stability signal acquisition. Vibration monitoring sensor 2 13, located at the bearing housing of spindle 14, also acquires radial vibration data of spindle 14 at a frequency of 500Hz. By monitoring this position signal, the system can isolate the background vibration caused by the imbalance of the transmission chain itself, and help determine the impact of tool wear on the overall rotational stability.

[0053] S203. Acquisition of transmission link attenuation characteristics. Vibration monitoring sensor 315, located inside the transmission housing, collects far-field vibration signals to establish a model of vibration energy transmission in the mechanical structure and to serve as a reference for environmental noise.

[0054] S204. Construction of the three-dimensional vibration vector. The vibration signal acquisition module 100 uses the CAN bus clock synchronization protocol to aggregate the signals acquired by the three position sensors. The system ensures that the sampling time difference of each signal is ≤0.1ms, thereby constructing the axial, radial, and circumferential components at the same moment into a time-synchronized three-dimensional vibration feature vector, ensuring signal time consistency to support subsequent multi-dimensional feature fusion.

[0055] S205. Active damping adjustment based on vibration feedback. The data processing and early warning module 300 calculates the current vibration energy level based on the real-time three-dimensional vibration vector. When an upward trend in vibration amplitude is detected, the control system sends a gain command to the magnetorheological damper 9 to increase the current intensity in the excitation coil. The shear yield stress of the magnetorheological fluid increases with the increase of the magnetic field, which increases the damping force of the damper and suppresses the vibration of the tool.

[0056] S206. Physical Coupling of Mechanical Load and Current Signal. When the magnetorheological damper 9 performs vibration damping operation, the non-stationary random load generated by the tool cutting hard rock is transmitted to the damper shaft 6. Due to the axial reciprocating motion of the damper shaft 6, the drastic fluctuations in the external mechanical load cause changes in the air gap magnetic flux of the internal magnetic circuit of the damper, which in turn causes dynamic disturbances in the coil inductance. This mechanical load fluctuation is directly manifested as subtle modulation and fluctuations in the current waveform in the excitation circuit.

[0057] S207. High-frequency characteristic capture of excitation current. The current monitoring unit 11 synchronously acquires the output current of the magnetic field generator and records the real-time amplitude, frequency, and instantaneous fluctuation curve of the current. This unit captures the high-frequency fluctuation component superimposed on the reference excitation current, thereby indirectly obtaining the cutting load characteristics by utilizing the electrical response of the magnetorheological damper 9 without adding an additional force sensor.

[0058] S208. High-temperature and high-pressure protection of the main seal. During the data acquisition process, the perfluoroelastomer directional lip seal 2 utilizes the material's high-temperature and corrosion-resistant properties to prevent the magnetorheological fluid from leaking out under the influence of thermal expansion and pressure fluctuations. The accompanying wave spring 3 provides a constant axial preload, compensating for the gaps caused by seal wear, maintaining the stability of the working pressure inside the cavity, and ensuring that the current signal reflects the actual load changes rather than pressure imbalances caused by medium leakage.

[0059] S209. Physical barrier against external dust. The PTFE-coated carbon fiber V-ring 4 utilizes its low coefficient of friction and high wear resistance to scrape away rock powder adhering to the shaft surface as the damper shaft 6 moves. This structure effectively blocks the entry of external micro-dust, preventing dust from mixing into the magnetorheological fluid and causing particle agglomeration, thus ensuring the consistency of the magnetorheological fluid's rheological properties.

[0060] S210, Lubrication and Adsorption of the Buffer Chamber. The silicon-based grease containing nano-ceramic particles, which is placed in the annular buffer chamber 5, reduces the frictional resistance of the damper shaft 6 on the one hand, and adsorbs the trace dust that may break through the outer seal on the other hand, ensuring the long-term stability of the internal electromagnetic characteristics of the magnetorheological damper 9.

[0061] After acquiring the raw physical signals in step S2, in order to eliminate interference from industrial environmental noise and ensure strict correspondence between multi-source signals in the time dimension, the system enters the data transmission and signal conditioning stage in step S3. The specific sub-steps of this stage are as follows:

[0062] S301. Hardware-level analog signal conditioning. Before analog-to-digital conversion, the vibration signal acquisition module 100 and the current monitoring unit 11 first perform hardware conditioning on the original analog signal. For the high-impedance charge signal output by the piezoelectric sensor, the signal conditioning circuit uses a charge amplifier to convert it into a low-impedance voltage signal and performs impedance matching. At the same time, a second-order Butterworth low-pass filter is set as an anti-aliasing filter. Based on the aforementioned sampling frequency of 500Hz, its cutoff frequency is set to 250Hz to filter out high-frequency noise components higher than the Nyquist frequency and prevent aliasing distortion from occurring during signal discretization. The specific design of the above-mentioned charge amplification and hardware filtering circuit can be implemented by those skilled in the art according to general electronic design manuals and is well-known in the field, so it will not be described in detail here.

[0063] S302. Denoising and Cleaning of Digital Signals. The digital signals after A / D conversion are preprocessed by the microprocessors at each acquisition node. To address the strong electromagnetic interference present in the working environment of hard rock tunneling machines, the system employs a digital notch filter with a center frequency set to 50Hz and its multiples to eliminate power line interference. Simultaneously, to address the sudden electromagnetic pulse noise generated by the start-up and shutdown of large motors, the system applies a mean-mode filtering algorithm or a wavelet threshold denoising algorithm to identify and remove isolated outliers in the signal, retaining the true trend terms of the vibration and current signals to ensure that the signal-to-noise ratio of subsequent feature extraction meets the model input requirements.

[0064] S303. CAN bus-based timestamp synchronization mechanism. The conditioned digital signal is transmitted to the data processing and early warning module 300 via the CAN bus. To address the discrepancy caused by the asynchronous physical clock sources between the vibration acquisition module and the magnetorheological monitoring module, the system employs a timestamp-based synchronization transmission protocol. The data processing and early warning module 300, acting as the master node, periodically broadcasts a global synchronization clock frame. The vibration signal acquisition module 100 and the current monitoring unit 11, acting as slave nodes, calibrate their local clocks upon receiving the synchronization frame and embed a timestamp based on a unified time base into the header of each uploaded data packet.

[0065] S304. Timing Alignment of Multi-Source Heterogeneous Data. After receiving CAN data frames from different nodes, the data processing and early warning module 300 parses the timestamp information of each data packet. Using the time axis of the vibration signal as a reference, the system performs time compensation on the current signal using an interpolation algorithm, mapping the current signal vector to the time point of the vibration signal. Through this mechanism, the deviation between the impact peak value in the vibration waveform and the load fluctuation characteristics in the current waveform on the time axis is controlled within 0.1ms, providing a precise timing reference for subsequent data-level or feature-level fusion of mechanical and electromagnetic domain signals.

[0066] After completing the temporal alignment and preprocessing of multi-source heterogeneous data in step S3, the system enters the multi-dimensional feature fusion and intelligent discrimination stage in step S4. In this stage, the data processing and early warning module 300 executes the following sub-steps:

[0067] S401. Extraction of Time-Domain Statistical Features in the Mechanical Domain. The data processing and early warning module 300 performs statistical analysis on the three-dimensional vibration signal within a preset time window. First, the maximum absolute value of the signal within the time window is extracted as the vibration peak value, which corresponds to the maximum transient impact intensity when the tool cuts the rock. Second, the root mean square (RMS) value is calculated, representing the average energy level of the vibration signal over a period of time. Finally, the peak factor, i.e., the ratio of the peak value to the RMS value, is calculated. These feature parameters are used to capture wear information throughout the entire process, from early weak impacts to later severe friction.

[0068] S402. Extraction of Electromagnetic Domain Load Mapping Features. For the synchronously acquired excitation current signal, the system extracts electromagnetic domain features reflecting the fluctuation characteristics of the cutting load. The system calculates the current amplitude fluctuation coefficient, which is the ratio of the standard deviation to the arithmetic mean of the excitation current signal within a time window, used to quantify the degree of magnetoelectric coupling modulation caused by mechanical load fluctuations. Simultaneously, the system extracts the frequency deviation value through spectrum analysis, i.e., the difference between the actual main frequency of the current signal and the reference excitation frequency. This value reflects the phase lag and frequency traction effect of the magnetorheological vibration damper 9 when responding to severe forced vibration.

[0069] S403. Construction and Model Loading of Multidimensional Heterogeneous Feature Vectors. The system serially concatenates and normalizes the extracted mechanical and electromagnetic domain features to construct a joint feature vector. Subsequently, the system calls a pre-set tool condition discrimination model. This model is a four-class classification model built based on the Support Vector Machine (SVM) algorithm, and its training data comes from laboratory simulations of tunnel hard rock conditions and actual field operation data. The model classifies tool conditions into new tools, light wear (wear amount 0.5-1mm), moderate wear (wear amount 1-2mm), and heavy wear (wear amount >2mm). During the model construction process, 10,000 sets of sample data containing three-dimensional vibration features and current fluctuation features were used. After normalization, the training set and test set were divided in a 7:3 ratio, and the penalty parameters were optimized using a grid search method. Finally, 5-fold cross-validation ensured that the accuracy of the model on the test set was ≥95%, and the accuracy in actual field application was ≥92%.

[0070] S404. Real-time Status Determination Based on High-Dimensional Mapping. In real-time monitoring, the SVM model uses the Gaussian radial basis function (RBF) to map the joint feature vector of the current input to a high-dimensional space. The model uses the optimal classification hyperplane determined during training to calculate the geometric interval between the input sample points and each decision boundary, determining the specific wear category of the tool (new tool, light, moderate, or heavy). This process utilizes the sensitivity of vibration features to micro-scraping and the responsiveness of current features to macroscopic resistance to achieve accurate identification of the tool's health status.

[0071] S405. Setting the Basic Alarm Threshold. In addition to model discrimination, the system performs threshold monitoring in parallel. Based on the tool's full life cycle test data and statistical analysis of the 95% confidence interval, the system sets the basic alarm thresholds as follows: vibration peak value ≤ 5g, frequency ≤ 500Hz, and current fluctuation coefficient ≤ ±5%. This basic threshold corresponds to the critical signal characteristics when the tool fails with a critical wear amount (>2mm).

[0072] S406. Adaptive Dynamic Threshold Calibration. To address false alarms caused by fixed thresholds under varying geological conditions, the system reads the rock hardness, tunneling speed, and cutting load parameters of the tunnel boring machine at a frequency of 1Hz. The system dynamically corrects the base thresholds based on a pre-trained mapping model, adaptable to working conditions ranging from 25-90MPa hard rock hardness, 15-60mm / r tunneling speed, and 80-200kN cutting load. The specific dynamic correction rule is as follows: based on the currently monitored rock hardness, for every 10MPa increase in rock hardness, the system automatically raises the vibration peak threshold by 8% and simultaneously increases the current fluctuation threshold by 6%. Through this mechanism, the system automatically relaxes the alarm upper limit in hard rock areas to accommodate high cutting resistance, and automatically tightens the threshold in soft rock areas to maintain high sensitivity to early damage.

[0073] Based on the tool health status classification results and dynamic threshold determination conclusions obtained in step S4, the system enters the hierarchical early warning and closed-loop feedback stage in step S5. The specific implementation of this stage is as follows:

[0074] S501. Construction of Comprehensive Decision Logic. The data processing and early warning module 300 uses a "model-threshold dual verification" mechanism to generate the final early warning command. The system compares the classification results of the SVM model (new tool, light wear, moderate wear, heavy wear) with the status of the real-time signal relative to the dynamic threshold in real time. According to the settings of this embodiment, when the SVM model determines the result as "moderate wear" and the real-time signal characteristic parameters do not exceed the dynamic threshold, the system determines that a first-level alarm condition is triggered; when the SVM model determines the result as "heavy wear", or the amplitude of the real-time vibration / current signal exceeds the working condition adaptive dynamic alarm threshold calculated in step S4, the system determines that a second-level alarm condition is triggered.

[0075] S502, Graded Early Warning and Shutdown Protection Execution. The early warning execution unit executes differentiated audible and visual alarm actions based on the above judgment results. If a Level 1 alarm is triggered, the unit controls the yellow indicator light to flash and emits an intermittent alarm sound to alert the operator. At this time, the tunneling machine continues to run, but the system automatically increases the monitoring frequency. If a Level 2 alarm is triggered, the unit controls the red indicator light to remain on and emits a continuous alarm sound, indicating that the cutter wear is accelerating or is about to fail. Simultaneously with the triggering of a Level 2 alarm, the system outputs a shutdown warning signal to the main control PLC of the hard rock tunneling machine via the CAN bus, recommending that the operator stop the machine in time to replace the cutters. For extremely dangerous situations (such as signals far exceeding the threshold), the main control PLC can be configured to automatically execute emergency stop logic, that is, cut off the power supply to the drive motor and control the retraction of the propulsion hydraulic cylinder.

[0076] S503. Closed-loop damping adjustment based on state feedback. In addition to alarm actions, the system also performs closed-loop control of the magnetorheological damper 9 based on the determined state. When the determination result is in the first-level alarm state (moderate wear) or the model shows slight wear even without an alarm, the data processing module outputs an enhanced damping control signal to the magnetorheological damping monitoring module. The system superimposes a preset gain current on the set value of the basic excitation current to increase the excitation current and improve the shear yield stress of the magnetorheological fluid. By actively increasing the damping force, the system can suppress the vibration deterioration caused by tool wear, thereby delaying the wear process and optimizing the cutting conditions to a certain extent.

[0077] S504. Verification of vibration reduction optimization under operating conditions. The current monitoring unit 11 continuously captures load feedback after adjusting the excitation current. If it is found that the vibration amplitude and current fluctuation coefficient both show a downward trend after increasing the excitation current, the system confirms that the closed-loop control is effective and maintains the current enhancement parameters; if the signal does not improve significantly after adjustment, the system will maintain the current level two alarm status and prompt to check the performance of the magnetorheological vibration damper 9.

Claims

1. An integrated condition monitoring, early warning, and vibration reduction system for hard rock tunnel boring machine cutters, characterized in that, include: The vibration signal acquisition module (100) is located at the key mechanical nodes of the hard rock tunneling machine to capture the mechanical response generated during the operation of the cutting tool in real time and output multi-dimensional vibration signals. A magnetorheological vibration damping monitoring module (200) is installed on the vibration transmission path and includes a magnetorheological vibration damper (9) and a current monitoring unit (11). The magnetorheological vibration damper (9) adjusts the excitation current in response to external control commands to output a variable damping force. The current monitoring unit (11) synchronously captures the output excitation current signal generated by the excitation coil of the magnetorheological vibration damper (9) when it is subjected to mechanical load. The data processing and early warning module (300) establishes a time-synchronized communication connection with the vibration signal acquisition module (100) and the magnetorheological vibration reduction monitoring module (200) via the CAN bus; The data processing and early warning module (300) performs feature fusion between the received multidimensional vibration signal and the output excitation current signal to determine the tool wear state, and sends a graded early warning command to the external alarm device or a shutdown command to the main control system of the hard rock tunneling machine according to the determination result. The data processing and early warning module (300) performs feature extraction, including: The mechanical domain features of the multidimensional vibration signal are extracted. The mechanical domain features include the maximum absolute value of the signal within a preset time window as the vibration peak value, the root mean square value of the signal, and the ratio of the vibration peak value to the root mean square value as the peak factor. The electromagnetic domain features of the output excitation current signal are extracted. The electromagnetic domain features include the ratio of the standard deviation to the arithmetic mean of the signal within a preset time window as the current amplitude fluctuation coefficient, and the difference between the actual main frequency of the signal and the reference excitation frequency as the frequency deviation value. The specific method for determining the execution status of the data processing and early warning module (300) is as follows: The mechanical domain features and the electromagnetic domain features are concatenated and normalized to construct a joint feature vector; The joint feature vector is input into a preset tool state discrimination model. The tool state discrimination model adopts the support vector machine algorithm, uses the Gaussian radial basis kernel function to map the joint feature vector to a high-dimensional space, solves the geometric interval between the sample points and the decision boundary, and outputs the tool wear state category. The wear state category includes new tool, light wear, moderate wear and heavy wear.

2. The integrated condition monitoring, early warning, and vibration reduction system for hard rock tunnel boring machine cutters according to claim 1, characterized in that, The vibration signal acquisition module (100) includes multiple piezoelectric sensors distributed at the mechanical nodes of the hard rock tunneling machine: Vibration monitoring sensor 1 (8) is rigidly installed at the front end of the cutter mounting seat of the tunneling head (7) of the hard rock tunneling machine, and is used to collect the three-dimensional spatial vibration signal of the cutter. Vibration monitoring sensor 2 (13) is installed at the bearing seat of the main shaft (14) of the hard rock tunneling machine to collect radial vibration data of the main shaft (14); Vibration monitoring sensor three (15) is installed on the inner wall of the transmission box housing of the hard rock tunneling machine to collect far-field vibration signals.

3. The integrated status monitoring, early warning, and vibration reduction system for hard rock tunnel boring machine cutters according to claim 1, characterized in that, The magnetorheological damper (9) adopts a three-stage sealing structure, including: The magnetorheological cylinder barrel (1) has a perfluoroelastomer directional lip seal (2) and a polytetrafluoroethylene-coated carbon fiber V-ring (4) arranged sequentially from the inside to the outside at its port. The wave spring (3), used in conjunction with the perfluoroelastomer directional lip seal (2), provides mechanical preload; An annular buffer cavity (5) is located between the perfluoroelastomer directional lip seal (2) and the polytetrafluoroethylene-coated carbon fiber V-ring (4), and is filled with silicone-based grease containing nano-ceramic particles. The damper shaft (6) passes through the three-stage sealing structure and reciprocates within the magnetorheological cylinder (1).

4. The integrated condition monitoring, early warning, and vibration reduction system for hard rock tunnel boring machine cutters according to claim 2, characterized in that, The data processing and early warning module (300) includes: The data acquisition system (10) has a built-in signal conditioning circuit and an industrial control computer. It synchronously receives the vibration signals collected by the vibration monitoring sensor 1 (8), the vibration monitoring sensor 2 (13), and the vibration monitoring sensor 3 (15) and the current signal collected by the current monitoring unit (11) via the CAN bus. An audible and visual alarm (12) is connected to the data acquisition system (10) and configured to issue a graded alarm when an abnormal command is received.

5. The integrated condition monitoring, early warning, and vibration reduction system for hard rock tunnel boring machine cutters according to claim 4, characterized in that, The specific method by which the data acquisition system (10) performs data synchronization is as follows: Broadcasts a global synchronization clock frame as the master node; The vibration signal acquisition module (100) and the current monitoring unit (11) act as slave nodes. After receiving the synchronization frame, they calibrate the local clock and embed the acquisition time timestamp based on a unified time base in the header of the data packet uploaded in each frame. The data acquisition system (10) parses the timestamp, takes the time axis of the multidimensional vibration signal as a reference, uses an interpolation algorithm to perform time compensation on the output excitation current signal, and maps the output excitation current signal to the time point of the multidimensional vibration signal.

6. The integrated status monitoring, early warning, and vibration reduction system for hard rock tunnel boring machine cutters according to claim 4, characterized in that, The data processing and early warning module (300) is equipped with a dynamic threshold calibration function that adapts to operating conditions: Read the current rock hardness parameters of the hard rock tunneling machine; The basic alarm threshold is corrected using a mapping model. The correction rule is as follows: take the currently monitored rock hardness as the benchmark. If the rock hardness increases, the threshold of the vibration peak value is increased proportionally, and the threshold of the current amplitude fluctuation coefficient is increased proportionally.

7. The integrated status monitoring, early warning, and vibration reduction system for hard rock tunnel boring machine cutters according to claim 6, characterized in that, The hierarchical early warning logic executed by the data processing and early warning module (300) is as follows: Combining the wear status category of the tool with the corrected basic alarm threshold, when the tool status discrimination model determines that the wear is moderate and the real-time signal characteristic parameters do not exceed the corrected basic alarm threshold, it is determined to be a level one alarm condition, and the yellow indicator light of the sound and light alarm (12) is controlled to flash and emit an intermittent alarm sound. When the tool condition discrimination model determines that the tool is severely worn, or when the real-time signal characteristic parameters exceed the corrected basic alarm threshold, it is determined to be a level two alarm condition. The red indicator light of the sound and light alarm (12) is kept on and a continuous alarm sound is emitted. At the same time, a stop signal is output to the main control PLC of the hard rock tunneling machine.

8. A method for integrated condition monitoring, early warning, and vibration reduction of hard rock tunnel boring machine cutters, employing the integrated condition monitoring, early warning, and vibration reduction system for hard rock tunnel boring machine cutters as described in any one of claims 1-7, characterized in that... Includes the following steps: After the system is started with the hard rock tunneling machine, the data processing and early warning module (300) loads the preset tool state discrimination model and threshold parameters, the magnetorheological vibration reduction monitoring module (200) enters the working state, and each sensor executes the self-test program; During the tunneling operation, the vibration signal acquisition module (100) collects the multi-dimensional vibration signal of the cutting tool in real time, and the magnetorheological damper (9) attenuates the tunneling vibration. At the same time, the current monitoring unit (11) collects the output excitation current signal of the magnetorheological damper (9) when it is working. The acquired multidimensional vibration signal, after being filtered and amplified, is synchronously transmitted to the data processing and early warning module (300) via the CAN bus along with the output excitation current signal, and the signal is processed to remove power frequency interference and pulse noise. The data processing and early warning module (300) extracts the peak value, root mean square, peak factor of the multidimensional vibration signal and the amplitude fluctuation coefficient and frequency deviation value of the output excitation current signal, respectively, and inputs the extracted feature parameters into the tool state discrimination model for calculation to determine the current wear state of the tool. Based on the judgment results, a graded early warning strategy is implemented. If the signal does not exceed the dynamic threshold but the model determines that the wear is moderate, a level one audible and visual alarm is triggered. If the signal exceeds the dynamic threshold or the model determines that the wear is severe, a level two audible and visual alarm will be triggered and a shutdown command will be issued to protect the equipment.

Images