Intelligent anti-seismic system and method for cross-fault tunnel

Abstract

The application discloses an intelligent anti-seismic system and method for a cross-fault tunnel, and the system comprises an anti-seismic integrated support body, a multi-source sensing array, an embedded processing unit, a stress-damage collaborative execution controller and a dual-mode redundant transmission and storage module; the method comprises the following steps: the multi-source sensing array synchronously collects data of a cross-fault region; denoising filtering and feature preprocessing are performed by a hardware computing unit; stress and strain calculation, energy dissipation analysis and structure damage assessment are completed, and optimal anti-seismic control instructions are output; the stress-damage collaborative execution controller receives the control instructions, and adjusts the structure parameters of the anti-seismic integrated support body; the control parameters are fine-tuned; incremental training is started according to a set period to optimize model parameters, and redundant switching is automatically triggered when a fault is detected. The application realizes direct conversion from a theoretical model to engineering application, and the optimal parameters deduced by theory can be adapted to engineering actual working conditions in real time.

Description

Technical Field

[0001] This invention belongs to the field of tunnel engineering vibration reduction technology, specifically relating to an intelligent vibration reduction system and method for cross-fault tunnels. Background Technology

[0002] As a key node in transportation infrastructure, cross-fault tunnels need to directly cross or be adjacent to active faults. Under seismic loads, fault displacement and seismic wave propagation can easily cause serious disasters such as tunnel lining cracking, joint deformation, and surrounding rock instability, and may even lead to the overall failure of the tunnel. Therefore, seismic design is the core technical challenge of cross-fault tunnel engineering.

[0003] At present, the anti-vibration technology for cross-fault tunnels mainly revolves around passive protection and semi-active control. Although it has achieved certain application results, the following prominent problems still exist: the theoretical results are disconnected from engineering implementation, the damage assessment and control accuracy is insufficient, the system reliability and disaster resistance are weak, and the intelligence and operation and maintenance adaptability are poor.

[0004] Therefore, there is an urgent need to develop an intelligent vibration reduction method for cross-fault tunnels. Summary of the Invention

[0005] This invention is proposed to solve the problems existing in the prior art, and its purpose is to provide an intelligent anti-vibration and reduction system and method for cross-fault tunnels.

[0006] The technical solution of this invention is: an intelligent anti-vibration and damping system for cross-fault tunnels, comprising: The integrated shock absorption support body includes segmented lining, three-layer flexible joints, and rubber concrete shock absorption layer; A multi-source sensor array is used to simultaneously acquire geological parameters, seismic response data, structural damage data, and environmental data across fault zones. The embedded processing unit integrates a hardware computing unit and an inference core to preprocess the acquired data and output anti-vibration control commands. The stress-damage co-execution controller is communicatively connected to the embedded processing unit to receive the anti-vibration control command and dynamically adjust the structural parameters of the integrated anti-vibration support body based on the three closed-loop control formula. The dual-mode redundant transmission and storage module is used to ensure stable data transmission and secure storage, and supports automatic switching and data retransmission in case of link failure.

[0007] Furthermore, the intelligent shock absorption system also includes an early warning and safety protection module, which is used to monitor the system's operating status in real time and trigger early warning and emergency protection when abnormalities occur.

[0008] Furthermore, the embedded processing unit performs logical execution of the CDP constitutive model, Rayleigh damping calculation, multi-dimensional damage assessment formula, three-closed-loop control formula, and dual-objective optimization function.

[0009] Furthermore, the system includes geological parameter sensors, seismic response sensors, structural damage sensors, and environmental adaptation sensors. These sensors support synchronous sampling, and the sampled data includes device identification, section number, monitoring location, timestamp, data value, and checksum.

[0010] Furthermore, the stress-damage co-execution controller incorporates a three-loop control formula, which includes a position loop, a stress loop, and a damage loop.

[0011] Furthermore, the dual-mode redundant transmission and storage module includes a dual-mode redundant transmission unit and a hierarchical redundant storage unit. The dual-mode redundant transmission unit includes a wired transmission link and a wireless transmission link, and automatically switches to the wireless transmission link when the wired transmission link is interrupted. The hierarchical redundant storage unit uses a multi-layer architecture to store data and supports hardware write protection and encrypted storage.

[0012] Furthermore, the embedded processing unit automatically initiates incremental training at a set cycle, fine-tunes model parameters based on accumulated monitoring data and execution feedback data, and makes real-time decisions without interruption during the training process.

[0013] Furthermore, the segmented lining is cast with high-strength material and has reserved hardware installation slots, and the three-layer flexible joint is composed of multiple layers of heterogeneous materials.

[0014] A method for an intelligent vibration damping system for cross-fault tunnels includes the following steps: A. Geological parameters, seismic response data, structural damage data, and environmental data of the cross-fault area are collected synchronously through a multi-source sensor array; B. The collected data is transmitted to the embedded processing unit, where noise reduction filtering and feature preprocessing are performed by the hardware computing unit; C. By calling the solidified core mechanical formulas and intelligent models, stress and strain calculations, energy dissipation analysis, and structural damage assessment are completed, and the optimal anti-vibration control commands are output. D. The control command is received by the stress-damage co-execution controller, and the structural parameters of the integrated anti-vibration and damping support body are dynamically adjusted based on the three closed-loop control formula; E. Collect the adjusted feedback data in real time through the sensor array, and fine-tune the control parameters based on the feedback; F. Continuously accumulate data and start incremental training at set intervals to optimize model parameters, and automatically trigger redundancy switching when a fault is detected.

[0015] The beneficial effects of this invention are as follows: This invention enables the direct transformation from theoretical models to engineering applications, and the optimal parameters derived from the theory can be adapted to actual engineering conditions in real time.

[0016] This invention employs a multi-dimensional damage assessment formula and a three-closed-loop control formula, comprehensively considering various damage indicators such as stress, strain, and cracks, to achieve a balance between vibration reduction effect and structural safety.

[0017] This invention supports incremental iterative optimization of the model, which can adapt to dynamic changes in geological conditions. It also supports remote monitoring and troubleshooting, reducing operation and maintenance costs. Attached Figure Description

[0018] Figure 1 This is a flowchart illustrating the overall architecture of the present invention; Figure 2 This is a detailed flowchart of the initial deployment of this invention; Figure 3 This is a detailed flowchart of the intelligent operation closed loop of the present invention; Figure 4 This is a flowchart of the redundancy protection and iterative refinement process of this invention. Detailed Implementation

[0019] The present invention will now be described in detail with reference to the accompanying drawings and embodiments: like Figures 1 to 4 As shown, an intelligent vibration damping system for cross-fault tunnels includes: The integrated shock absorption support body includes segmented lining, three-layer flexible joints, and rubber concrete shock absorption layer; A multi-source sensor array is used to simultaneously acquire geological parameters, seismic response data, structural damage data, and environmental data across fault zones. The embedded processing unit integrates a hardware computing unit and an inference core to preprocess the acquired data and output anti-vibration control commands. The stress-damage co-execution controller is communicatively connected to the embedded processing unit to receive the anti-vibration control command and dynamically adjust the structural parameters of the integrated anti-vibration support body based on the three closed-loop control formula. The dual-mode redundant transmission and storage module is used to ensure stable data transmission and secure storage, and supports automatic switching and data retransmission in case of link failure.

[0020] The intelligent shock absorption system also includes an early warning and safety protection module, which is used to monitor the system's operating status in real time and trigger early warning and emergency protection when abnormalities occur.

[0021] The embedded processing unit performs logical execution of the CDP constitutive model, Rayleigh damping calculation, multi-dimensional damage assessment formula, three-closed-loop control formula, and bi-objective optimization function.

[0022] The sensors include geological parameter sensors, seismic response sensors, structural damage sensors, and environmental adaptation sensors. These sensors support synchronous sampling, and the sampled data includes device identification, section number, monitoring location, timestamp, data value, and checksum.

[0023] The stress-damage co-execution controller incorporates a three-loop control formula, which includes a position loop, a stress loop, and a damage loop.

[0024] The dual-mode redundant transmission and storage module includes a dual-mode redundant transmission unit and a hierarchical redundant storage unit. The dual-mode redundant transmission unit includes a wired transmission link and a wireless transmission link, and automatically switches to the wireless transmission link when the wired transmission link is interrupted. The hierarchical redundant storage unit uses a multi-layer architecture to store data and supports hardware write protection and encrypted storage.

[0025] The embedded processing unit automatically starts incremental training at a set period, fine-tunes the model parameters based on accumulated monitoring data and execution feedback data, and makes real-time decisions without interruption during the training process.

[0026] The segmented lining is cast with high-strength material and has reserved hardware installation slots. The three-layer flexible joint is composed of multiple layers of heterogeneous materials.

[0027] In the support body, the three-layer flexible joint is composed of multiple heterogeneous materials and is formed by bonding process. Its width and elastic modulus parameters are the optimal values ​​calibrated by the bi-objective optimization function, which can meet the dynamic adjustment requirements. Meanwhile, the rubber particle size and pouring thickness of the rubber concrete damping layer are adapted to the coverage requirements of the fault stress concentration area, and the bonding strength with the initial support and secondary lining meets the engineering safety requirements.

[0028] In the hardware structure of this system, the embedded processing unit can be, but is not limited to, an embedded AI chip. The embedded AI chip adopts a heterogeneous architecture, integrating a high-performance computing core, an FPGA hardware unit, and an AI inference module. It supports the fixed operation of core formulas and models such as the CDP constitutive model formula and the Rayleigh damping calculation formula. It has low power consumption, low latency, and is suitable for deployment in the confined space of tunnels.

[0029] A method for an intelligent vibration damping system for cross-fault tunnels includes the following steps: A. Geological parameters, seismic response data, structural damage data, and environmental data of the cross-fault area are collected synchronously through a multi-source sensor array; B. The collected data is transmitted to the embedded processing unit, where noise reduction filtering and feature preprocessing are performed by the hardware computing unit; C. By calling the solidified core mechanical formulas and intelligent models, stress and strain calculations, energy dissipation analysis, and structural damage assessment are completed, and the optimal anti-vibration control commands are output. D. The control command is received by the stress-damage co-execution controller, and the structural parameters of the integrated anti-vibration and damping support body are dynamically adjusted based on the three closed-loop control formula; E. Collect the adjusted feedback data in real time through the sensor array, and fine-tune the control parameters based on the feedback; F. Continuously accumulate data and start incremental training at set intervals to optimize model parameters, and automatically trigger redundancy switching when a fault is detected.

[0030] Before monitoring, this method requires determining the monitoring section and deploying the sensor array. The specific process is as follows: First, through geological exploration, key parameters such as the location, width, and dip angle of the fault are determined, a tunnel geological model is established, and the Rayleigh damping calculation formula is called ( (where C is the damping matrix, M is the mass matrix, and K is the stiffness matrix) Calculate energy dissipation, determine the fault influence range, and delineate monitoring sections; Then, dense monitoring sections were deployed along the center and both sides of the fault fracture zone. Integrated sensing components were fixed at key parts such as the arch crown, arch shoulder, arch waist, arch foot, and invert arch of each section. Additional structural damage sensors were deployed to ensure that the collected data met the calculation requirements of the multidimensional damage assessment formula (SDEG). Next, routine monitoring sections were deployed in stable areas far from the fault. Basic sensing components were installed at key locations on each section, and the installation and fixing methods were the same as those for dense monitoring sections. Next, environmental adaptation sensors are deployed at the tunnel entrances and exits and the center of the fault to compensate for the impact of environmental factors on the sensor data and ensure the accuracy of the formula calculation. Finally, after all sensors are installed, the automatic calibration process is started upon power-on: using the built-in high-precision calibration component, the standard response value is compared with the initial acquired data, the deviation is calculated according to "deviation = (measured value - standard value) / standard value × 100%" and the sensor parameters are automatically corrected, with a short calibration time; a timed calibration cycle is set to ensure that the deviation between the acquired data and the simulation and test data meets the engineering accuracy requirements.

[0031] After the sensor array is deployed, the embedded processing unit is installed, as follows: First, the embedded AI chip and its supporting modules are fixed in the mounting slot reserved in the lining, and a shock-absorbing and buffering structure is configured to prevent vibration displacement. Then, the chip is connected to the backup computing unit through a high-speed bus to synchronize key formula parameters such as CDP constitutive model parameters, Rayleigh damping coefficients and model weights, and the synchronization delay meets the requirements. Then, connect to dual power supply and backup power to ensure continuous operation after power failure; Finally, a connection is established with the sensor array and actuator controller through the communication interface, and an adaptive communication protocol is configured to ensure that the data transmission latency meets the requirements.

[0032] The actuator controller and support body are adapted and installed as follows: First, the stress-damage co-execution controller is fixed on the reserved support on the tunnel sidewall and connected to the adjustment mechanism of the support body and the damping layer adapter components through pipelines and wiring harnesses; Then, the actuator stroke and parameter adjustment range were adjusted to ensure that the position and force control accuracy met the accuracy requirements of the three-closed-loop control formula; safety threshold parameters were set, and emergency stop response performance was tested. The position ring is: ; The stress ring is: ; The damage ring is: .

[0033] The deployment of transmission and storage is as follows: The dual-mode redundant transmission module deploys signal enhancement and relay equipment at intervals along the tunnel length to ensure full signal coverage; it is configured with automatic switching logic for link interruption, and the switching performance is tested to meet engineering requirements. The tiered redundant storage module is installed adjacent to the embedded AI chip and connected via a high-speed interface. It is configured with differentiated data storage strategies to ensure the secure storage of core formula parameters, monitoring data, and fault logs, and to test the normal read / write speed and encryption function of data.

[0034] The entire system will undergo integration testing, as detailed below: The entire system was powered on for testing. Simulated seismic wave signals were sent via the embedded AI chip, and the sensor array simultaneously collected multi-dimensional data and transmitted it to the chip. The chip then executed a series of core formula calculations through an FPGA hardware module. First, call the CDP constitutive model formula: ; Where σ is stress, ε is strain, and the rest are state parameters, the stress-strain relationship between the surrounding rock and the lining is calculated; Then, substitute the Rayleigh damping calculation to calculate the structural energy dissipation; ; Next, the damage status is assessed using the Multidimensional Damage Assessment Formula (SDEG). ; in, For the comprehensive damage value, These are stress, strain, and crack damage components, respectively. , where is the weighting coefficient; Then, based on the bi-objective optimization function, the optimal control command is output. ; Where R is the vibration damping effect index, D is the structural damage index, and P is the energy consumption control index; Finally, after receiving the command, the controller drives the actuator to adjust the support parameters. Feedback data shows that the vibration reduction effect meets the design requirements. The fault switching performance is tested, and the backup unit can quickly take over the service. According to the data consistency verification formula, the decision accuracy meets the engineering reliability requirements. ; in, Main unit output, Output for backup unit is the confidence coefficient.

[0035] After the re-integration and debugging, the system operation process is as follows: First, during the data acquisition phase, the multi-source sensor array synchronously collects data at a set sampling frequency, including geological parameters, seismic response data such as acceleration, strain, and shear stress, structural damage data, and environmental data. All data is transmitted to the embedded AI chip through an adapted communication protocol. The data frame contains fields such as device identifier, section number, monitoring location, timestamp, data value, and check code to ensure traceability. Then, data preprocessing and inference optimization preprocessing are performed: Embedded AI chips use FPGA hardware units to process multi-source data in parallel, performing noise reduction filtering and feature fusion algorithms. The db4 wavelet denoising method is fused with Kalman filtering, and the formula is as follows: , in, For the raw sensor data, The preceding covariance matrix is ​​used to improve the data signal-to-noise ratio, and the preprocessing delay meets real-time requirements. During inference optimization: the chip calls the solidified core mechanical formula and the improved hybrid intelligent model to complete the prediction of structural dynamic response, energy dissipation calculation, damage status assessment, and output the optimal support parameter command; the main and backup computing units issue commands after verification according to the consistency verification formula, and the inference delay meets the real-time requirements of the project. Next, during the feedback phase, after receiving the instruction, the stress-damage co-execution controller drives the actuator through a three-loop control formula to adjust the key parameters of the support body. The sensor array collects the adjusted response data in real time and feeds it back to the embedded AI chip. If the parameter deviation exceeds the set threshold, the chip triggers a secondary fine-tuning until the deviation meets the accuracy requirements. The three-loop control formula is the same as above. Finally, fault response and redundancy switching calculation redundancy: When a failure of the main computing unit is detected, the backup unit automatically takes over through real-time status monitoring signals, synchronizing all formula parameters, model weights and cached data to ensure uninterrupted calculation; transmission redundancy: if the wired transmission link is interrupted, the system automatically switches to the wireless link, and the switching time meets engineering requirements; when the network is down, the data is temporarily stored in the storage module, and automatically retransmitted according to data priority after the network is restored.

[0036] As operations and maintenance iterate... Routine operation and maintenance processes can be remotely monitored by viewing the operating status of each component, formula calculation results, and control command execution in real time through the backend system, without the need for on-site supervision. When the system reports a fault code, a self-diagnostic program can be initiated via remote commands to locate the fault and troubleshoot it. Software issues can be repaired by remote firmware upgrades, and hardware issues can be replaced on-site with hot-swappable technology, resulting in high replacement efficiency. Regular maintenance involves on-site inspection of hardware components according to a set schedule, cleaning impurities from sensor surfaces, tightening fixed structures, and testing protective performance to ensure adaptability to complex environments.

[0037] The incremental iteration of the model involves data accumulation through daily collection of sufficient homogeneous data, including monitoring data, formula calculation results, and execution feedback data, which are stored in the backup module. The embedded AI chip automatically initiates incremental training at a set cycle, fine-tuning only the key parameters of the model. The training time is short, and real-time decision-making is not interrupted. After training, the model's prediction accuracy for key parameters in the core formula is significantly improved. Remote upgrades are performed when enough geological condition change data is accumulated. The optimized model firmware is distributed through the backend and updated after security verification. Formula weights or parameters are adjusted synchronously to ensure that the model and formulas are adapted to the dynamic geological environment and meet the requirements for long-term service.

[0038] This invention enables the direct transformation from theoretical models to engineering applications, and the optimal parameters derived from the theory can be adapted to actual engineering conditions in real time.

[0039] This invention employs a multi-dimensional damage assessment formula and a three-closed-loop control formula, comprehensively considering various damage indicators such as stress, strain, and cracks, to achieve a balance between vibration reduction effect and structural safety.

[0040] This invention supports incremental iterative optimization of the model, which can adapt to dynamic changes in geological conditions. It also supports remote monitoring and troubleshooting, reducing operation and maintenance costs.

Claims

1. An intelligent vibration damping system for cross-fault tunnels, characterized in that: include: The integrated shock absorption support body includes segmented lining, three-layer flexible joints, and rubber concrete shock absorption layer; A multi-source sensor array is used to simultaneously acquire geological parameters, seismic response data, structural damage data, and environmental data across fault zones. The embedded processing unit integrates a hardware computing unit and an inference core to preprocess the acquired data and output anti-vibration control commands. The stress-damage co-execution controller is communicatively connected to the embedded processing unit to receive the anti-vibration control command and dynamically adjust the structural parameters of the integrated anti-vibration support body based on the three closed-loop control formula. The dual-mode redundant transmission and storage module is used to ensure stable data transmission and secure storage, and supports automatic switching and data retransmission in case of link failure.

2. The intelligent anti-vibration and damping system for cross-fault tunnels according to claim 1, characterized in that: The intelligent shock absorption system also includes an early warning and safety protection module, which is used to monitor the system's operating status in real time and trigger early warning and emergency protection when abnormalities occur.

3. The intelligent anti-vibration and damping system for cross-fault tunnels according to claim 1, characterized in that: The embedded processing unit performs logical execution of the CDP constitutive model, Rayleigh damping calculation, multi-dimensional damage assessment formula, three-closed-loop control formula, and bi-objective optimization function.

4. The intelligent anti-vibration and damping system for cross-fault tunnels according to claim 1, characterized in that: The sensors include geological parameter sensors, seismic response sensors, structural damage sensors, and environmental adaptation sensors. These sensors support synchronous sampling, and the sampled data includes device identification, section number, monitoring location, timestamp, data value, and checksum.

5. The intelligent anti-vibration and damping system for cross-fault tunnels according to claim 1, characterized in that: The stress-damage co-execution controller incorporates a three-loop control formula, which includes a position loop, a stress loop, and a damage loop.

6. The intelligent anti-vibration and damping system for cross-fault tunnels according to claim 1, characterized in that: The dual-mode redundant transmission and storage module includes a dual-mode redundant transmission unit and a hierarchical redundant storage unit. The dual-mode redundant transmission unit includes a wired transmission link and a wireless transmission link, and automatically switches to the wireless transmission link when the wired transmission link is interrupted. The hierarchical redundant storage unit uses a multi-layer architecture to store data and supports hardware write protection and encrypted storage.

7. The intelligent anti-vibration and damping system for cross-fault tunnels according to claim 1, characterized in that: The embedded processing unit automatically starts incremental training at a set period, fine-tunes the model parameters based on accumulated monitoring data and execution feedback data, and makes real-time decisions without interruption during the training process.

8. The intelligent anti-vibration and damping system for cross-fault tunnels according to claim 1, characterized in that: The segmented lining is cast with high-strength material and has reserved hardware installation slots. The three-layer flexible joint is composed of multiple layers of heterogeneous materials.

9. The method for an intelligent anti-vibration and damping system for a cross-fault tunnel according to claim 1, characterized in that: Includes the following steps: A. Geological parameters, seismic response data, structural damage data, and environmental data of the cross-fault area are collected synchronously through a multi-source sensor array; B. The collected data is transmitted to the embedded processing unit, where noise reduction filtering and feature preprocessing are performed by the hardware computing unit; C. By calling the solidified core mechanical formulas and intelligent models, stress and strain calculations, energy dissipation analysis, and structural damage assessment are completed, and the optimal anti-vibration control commands are output. D. The control command is received by the stress-damage co-execution controller, and the structural parameters of the integrated anti-vibration and damping support body are dynamically adjusted based on the three closed-loop control formula; E. Collect the adjusted feedback data in real time through the sensor array, and fine-tune the control parameters based on the feedback; F. Continuously accumulate data and start incremental training at set intervals to optimize model parameters, and automatically trigger redundancy switching when a fault is detected.

Images