Real-time dynamic optimization and adjustment method and system for tunneling process

Abstract

The invention discloses a real-time dynamic optimization adjustment method and system for a tunneling process, and relates to the technical field of intelligent tunneling, and the method comprises the following steps: obtaining the operation state data of tunneling equipment in real time, and recording the dynamic change characteristics of a stratum; the tunneling parameters are optimized and adjusted in real time, the tunneling machine is prevented from entering a resonance interval, fatigue damage of key components such as a cutter head and a thrust cylinder is reduced, the stability of equipment is improved, and the service life of the equipment is prolonged. A multi-objective optimization control strategy is adopted, and an intelligent algorithm is combined, so that the tunneling mode is automatically adjusted, the tunneling efficiency is improved, and manual intervention is reduced. According to the stratum prediction method based on machine learning, the adaptability of the heading machine to complex stratums is enhanced, stratum sudden change risks are warned in advance, and construction potential safety hazards are reduced. And meanwhile, an abnormal state emergency processing mechanism ensures that the tunneling process is stable and controllable, and the construction safety and reliability are improved.

Description

Technical Field

[0001] The present invention relates to the field of intelligent tunneling technology, and in particular to a real-time dynamic optimization and adjustment method and system for a tunneling process. Background Art

[0002] Real-time dynamic optimization and adjustment of the excavation process refers to the use of intelligent algorithms and control strategies to dynamically adjust key variables such as the operating mode, propulsion speed, cutterhead torque, and support parameters of the excavation equipment based on real-time collected environmental parameters, equipment status, and construction data during the excavation construction process to adapt to complex and changing geological conditions, improve excavation efficiency, reduce equipment wear and tear, and ensure construction safety.

[0003] This process typically relies on IoT sensors, data acquisition systems, and intelligent control systems. These systems continuously analyze factors such as ground changes ahead of the tunnel face, surrounding rock stability, and equipment load, enabling timely optimization of construction parameters. For example, during shield tunneling, soil pressure, grouting volume, and advance speed can be adjusted based on the geological conditions ahead to prevent ground collapse or cutterhead jamming. In mine tunneling, drilling and blasting parameters and ventilation strategies can be dynamically adjusted to improve construction efficiency and reduce safety risks.

[0004] Through real-time dynamic optimization and adjustment, not only can the excavation efficiency be improved, but also the construction cost can be reduced, the equipment failure rate can be reduced, and the construction safety can be guaranteed to the greatest extent.

[0005] Existing technologies have the following drawbacks: During the real-time dynamic optimization and adjustment of tunneling parameters, the lag in optimizing and adjusting tunneling parameters can trigger system resonance, leading to serious consequences. During shield machine or tunnel boring machine (TBM) construction, the adjustment of key parameters such as thrust, cutterhead speed, and support pressure typically relies on data analysis and feedback control. However, if the optimization system fails to fully consider the nonlinear response characteristics of the stratum, or if computational delays prevent timely updates of tunneling parameters, low-frequency resonances may form between the equipment and the stratum. This resonance can subject key components such as the cutterhead, thrust cylinder, and main bearing to periodic impacts beyond their design limits, accelerating equipment fatigue damage and even causing propulsion system instability. This can lead to machine jamming, cutterhead damage, or, in extreme cases, tunnel collapse, resulting in major safety incidents and economic losses. Therefore, during real-time optimization and adjustment of the tunneling process, the dynamic response of the stratum must be comprehensively considered to ensure the robustness and real-time performance of the adjustment strategy and avoid resonance hazards.

[0006] The above information disclosed in this Background section is only for enhancement of understanding of the background of the present disclosure and therefore it may contain information that does not form the prior art that is already known to a person of ordinary skill in the art. Summary of the Invention

[0007] The purpose of the present invention is to provide a real-time dynamic optimization and adjustment method and system for the tunneling process. By optimizing and adjusting the tunneling parameters in real time, the tunneling machine is prevented from entering the resonance range, fatigue damage to key components such as the cutterhead and propulsion cylinder is reduced, and the stability and service life of the equipment are improved. A multi-objective optimization control strategy is adopted in combination with an intelligent algorithm to achieve automatic adjustment of the tunneling mode, improve tunneling efficiency and reduce manual intervention. A stratum prediction method based on machine learning is used to enhance the adaptability of the tunneling machine to complex strata, provide early warning of the risk of stratum mutations, and reduce construction safety hazards. At the same time, the abnormal state emergency response mechanism ensures that the tunneling process is smooth and controllable, improves construction safety and reliability, and solves the problems in the above-mentioned background technology.

[0008] In order to achieve the above object, the present invention provides the following technical solution: a real-time dynamic optimization and adjustment method for a tunneling process, comprising the following steps:

[0009] Acquire the operating status data of tunneling equipment in real time and record the dynamic change characteristics of the stratum;

[0010] Based on the collected data, a nonlinear dynamic model of the tunneling process is constructed, and adaptive algorithms are used to identify the response characteristics of the formation and establish a mathematical relationship between the interaction between the equipment and the formation.

[0011] Using time series analysis combined with ground stiffness and roadheader vibration characteristics, the resonance threshold under the current tunneling state is calculated to determine whether there is a low-frequency resonance trend and to provide an early warning of potential instability risks.

[0012] Based on the prediction results, key tunneling parameters are dynamically adjusted, and the tunneling mode is optimized through adaptive control algorithms to avoid the resonance range of the equipment while ensuring tunneling efficiency.

[0013] Real-time monitoring of the impact of adjusted tunneling parameters on ground stability and equipment load. If resonance is still caused by parameter adjustments, further optimization of tunneling strategies is performed to form a closed-loop control system.

[0014] When unpredictable ground changes or abnormal equipment vibrations are detected, the emergency mechanism is automatically triggered to adjust the excavation rhythm, reduce the cutterhead load, and optimize the adjustment strategy based on historical data.

[0015] Preferably, the data collection step further includes multi-sensor data fusion technology to synchronously process tunneling parameters from different types of sensors to improve data accuracy and reliability;

[0016] Sensors include, but are not limited to, laser range sensors, inertial measurement units, formation pressure sensors, and accelerometers;

[0017] All data is first filtered and denoised after collection. The Kalman filter algorithm is used to reduce data fluctuations, and a time synchronization mechanism is used to ensure that data with different sampling frequencies can be compared and analyzed in the same time dimension.

[0018] In addition, in order to further improve the accuracy of stratum characteristic identification, the current excavation parameters are dynamically calibrated by combining historical excavation data and stratum geological exploration information to reduce data anomalies caused by stratum mutations, thereby ensuring that the dynamic optimization and adjustment of excavation parameters can be carried out on the basis of reliable data information.

[0019] Preferably, the nonlinear dynamic modeling step uses a deep neural network combined with a finite element analysis method to perform high-precision modeling of the interaction between the tunnel boring machine and the formation. The specific steps are as follows:

[0020] First, a physical model of the ground-roadhead interaction was constructed using the finite element analysis method to simulate the effects of different thrusts and cutterhead torques on ground disturbance.

[0021] Subsequently, a deep neural network was used for data training, enabling the model to learn the complex relationship between tunneling parameters and formation feedback from historical data and achieve adaptive adjustments.

[0022] In addition, genetic algorithms were further used to optimize the neural network parameters, so that the model can respond quickly when the formation changes and accurately predict the resonance critical value under different tunneling conditions.

[0023] Preferably, the resonance risk prediction step is not only based on traditional time series analysis methods, but also combines anomaly detection technology based on machine learning to more efficiently identify low-frequency resonances occurring during tunneling;

[0024] The vibration spectrum of the roadheader is analyzed in time series using an autoregressive integrated moving average model. The vibration characteristics are then classified using a support vector machine classification model to predict the resonance mode.

[0025] Automatically adjust the tunneling strategy when abnormal resonance trends are detected;

[0026] In addition, the prediction process uses the Bayesian update mechanism to continuously optimize the model parameters, so that the prediction accuracy continues to improve during the long-term operation.

[0027] Preferably, the resonance risk prediction step adopts a generalized resonance calculation model to calculate the resonance critical value through the following steps:

[0028] Calculate the main resonance frequency of the roadheader using the following expression:

[0029]

[0030] , where fr is the natural resonance frequency of the roadheader, k eq is the equivalent stiffness of the tunnel boring machine and the ground, m eq is the equivalent mass;

[0031] Calculate the damping characteristics of the formation, the calculation expression is as follows:

[0032]

[0033] , where ξ is the damping ratio, c eq is the equivalent damping coefficient;

[0034] Calculate the resonance gain factor of the system. The calculation expression is as follows:

[0035]

[0036] , where G r is the resonance gain factor, f is the actual vibration frequency of the roadheader;

[0037] To judge the resonance risk, the calculation expression is as follows:

[0038] G r >G th

[0039] , where G th is the resonance gain factor reference threshold;

[0040] If the resonance gain factor G r Exceeding the set threshold G th , a resonance warning is triggered and the tunneling parameters are adjusted to avoid the resonance risk.

[0041] Preferably, the specific steps of real-time optimization and adjustment are as follows:

[0042] Adopting multi-objective optimization control strategy to improve excavation efficiency while ensuring excavation stability;

[0043] The optimization strategy uses the particle swarm optimization algorithm to establish a nonlinear mapping relationship between key parameters and find the optimal adjustment solution in the shortest time;

[0044] In addition, the optimization strategy is combined with fuzzy control theory. When the ground conditions suddenly change, the operating parameters of the tunnel boring machine can be quickly adjusted, and through real-time feedback control, the tunneling process can always be in an efficient and safe working state.

[0045] Preferably, the real-time optimization and adjustment step uses a nonlinear time-varying optimization algorithm to calculate the optimal tunneling parameter adjustment value. The specific steps are as follows:

[0046] Construct a nonlinear constrained optimization problem, and the calculation expression is as follows:

[0047]

[0048] , where J is the objective function, P i is the current value of the ith tunneling parameter, is the optimal value of the i-th excavation parameter, w i is the weight factor, which indicates the importance of the i-th parameter to the objective function, and N is the total number of tunneling parameters that need to be optimized;

[0049] Define the constraints and the calculation expression is as follows:

[0050] P min ≤P i ≤P max

[0051] , where P min and P max are the minimum and maximum values of the tunneling parameters, respectively.

[0052] Preferably, the Lagrange multiplier method is used to solve the problem, and the calculation expression is as follows:

[0053]

[0054] , where is the Lagrange function, λ is the Lagrange multiplier;

[0055] Iteratively solve the optimal excavation parameters, and the calculation expression is as follows:

[0056]

[0057] , where is the updated excavation parameter, α is the step size factor, It is the partial derivative of the objective function, which indicates the direction of influence of the current parameters on the optimization target;

[0058] Adjust the tunnel boring machine parameters in real time to ensure the equipment operates in the optimal state and avoid resonance risks.

[0059] Preferably, the specific steps of dynamic feedback control are as follows:

[0060] Adopting an adaptive neuro-fuzzy inference system for control optimization, the control rules can be adjusted according to the changes in the ground and the status of the equipment during the tunneling process;

[0061] First, the historical excavation data is classified through the neural network self-learning mechanism, and the fuzzy control rules are adjusted according to the current excavation status;

[0062] Then, fuzzy logic reasoning is used to evaluate the operating status of the tunnel boring machine in real time to determine whether the current tunneling parameters are in the resonance risk range;

[0063] Finally, through fuzzy weight adjustment, the tunneling parameter adjustment strategy is continuously optimized to ensure the stability and efficiency of the tunneling system.

[0064] Preferably, the real-time dynamic optimization and adjustment system for the tunneling process includes a data acquisition and status monitoring module, a nonlinear dynamic modeling module, a resonance risk prediction module, an adaptive optimization and adjustment module, a closed-loop feedback control module, and an abnormal state emergency processing module:

[0065] The data acquisition and status monitoring module acquires the operating status data of the tunneling equipment in real time and records the dynamic change characteristics of the stratum;

[0066] The nonlinear dynamic modeling module builds a nonlinear dynamic model of the tunneling process based on the collected data, uses an adaptive algorithm to identify the response characteristics of the formation, and establishes a mathematical relationship between the interaction between the equipment and the formation;

[0067] The resonance risk prediction module uses a time series analysis method combined with the ground stiffness and vibration characteristics of the tunnel boring machine to calculate the resonance critical value under the current tunneling state, determine whether there is a low-frequency resonance trend, and warn of potential instability risks;

[0068] The adaptive optimization and adjustment module dynamically adjusts key tunneling parameters based on prediction results and optimizes the tunneling mode through adaptive control algorithms, allowing equipment operation to avoid resonance zones while ensuring tunneling efficiency.

[0069] A closed-loop feedback control module monitors in real time the impact of adjusted tunneling parameters on ground stability and equipment load. If the parameter adjustment still causes resonance, the tunneling strategy is further optimized to form a closed-loop control.

[0070] The abnormal state emergency processing module automatically triggers the emergency mechanism when unpredictable formation mutations or abnormal equipment vibrations are detected, adjusts the excavation rhythm, reduces the cutterhead load, and optimizes the adjustment strategy based on historical data.

[0071] In the above technical solution, the technical effects and advantages provided by the present invention are:

[0072] The present invention optimizes and adjusts the tunneling parameters in real time, combines nonlinear dynamic modeling and machine learning prediction, and effectively prevents the tunnel boring machine from entering the resonance range during the construction process. Traditional tunneling control methods mainly rely on manual experience or fixed parameter adjustment, and are unable to respond to the dynamic changes of the stratum in real time, resulting in a lag in the adjustment of the propulsion force, cutterhead speed and support pressure, which may trigger low-frequency resonance, causing fatigue damage to the cutterhead, uneven force on the propulsion cylinder, and even equipment instability. The present invention uses a generalized resonance calculation model (GRCM) and a nonlinear time-varying optimization algorithm (NTOA) to calculate the resonance critical value in real time, and uses closed-loop feedback control for dynamic adjustment to ensure that the tunneling parameters are always within a safe range. This method can significantly reduce the risk of damage to mechanical components caused by resonance, improve the long-term reliability of the tunnel boring machine, while reducing equipment maintenance costs and downtime, thereby improving construction efficiency.

[0073] The present invention adopts a multi-objective optimization control strategy to improve the excavation efficiency while ensuring the stable operation of the equipment. Traditional excavation methods usually adopt a fixed parameter operation mode. When encountering changes in the formation, manual intervention is often required to adjust the excavation parameters, resulting in limited construction efficiency, and even problems such as propulsion obstruction or cutterhead jamming when sudden changes in formation conditions occur. The present invention uses a particle swarm optimization (PSO) algorithm and a deep neural network (DNN) prediction model to automatically calculate the optimal combination of propulsion force, cutterhead speed and support pressure based on formation feedback data, and dynamically adjust the excavation mode to avoid the equipment running in an unnecessary inefficient state. In addition, the method can learn the optimal parameter control strategy in the excavation process in real time through an adaptive neural fuzzy inference system (ANFIS), so that the system has self-learning and self-optimization capabilities, reduces the need for manual intervention, and improves the level of construction automation, thereby maximizing the excavation speed while ensuring construction safety.

[0074] The present invention adopts a stratum prediction and optimization adjustment method based on machine learning to effectively improve the adaptability of the tunnel boring machine under different stratum conditions. When encountering sudden changes in strata (such as soft and hard alternating layers, fault fracture zones, water-rich sand layers, etc.), traditional tunneling methods are often unable to respond quickly, resulting in untimely adjustment of tunneling parameters, which may cause equipment resonance, stratum instability, shield posture deviation and other problems, seriously affecting construction safety. The present invention uses a support vector machine (SVM) classification algorithm and an autoregressive integrated moving average (ARIMA) model to monitor the changes in the mechanical properties of the stratum during tunneling in real time, predict the construction risks that may be brought about by sudden changes in strata in advance, and automatically adjust the tunneling parameters to ensure that the construction process is stable and controllable. In addition, the present invention adopts an abnormal state emergency response mechanism. When unpredictable stratum mutations or abnormal vibrations of equipment are detected, the system can automatically adjust the tunneling rhythm, reduce the cutterhead load, and optimize the adjustment strategy in combination with historical data, thereby minimizing construction safety hazards and improving the reliability of the tunneling system in complex stratum environments. BRIEF DESCRIPTION OF THE DRAWINGS

[0075] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, a brief introduction to the drawings required for use in the embodiments will be given below. Obviously, the drawings described below are only some embodiments recorded in the present invention. For ordinary technicians in this field, other drawings can also be obtained based on these drawings.

[0076] Figure 1 This is a flow chart of the method for real-time dynamic optimization and adjustment of the tunneling process of the present invention.

[0077] Figure 2 This is a module diagram of the real-time dynamic optimization and adjustment system for the tunneling process of the present invention. DETAILED DESCRIPTION

[0078] Example embodiments will now be described more fully with reference to the accompanying drawings. However, example embodiments can be implemented in many forms and should not be construed as limited to the examples set forth herein; rather, these example embodiments are provided so that the description of this disclosure will be thorough and complete and will fully convey the concepts of the example embodiments to those skilled in the art.

[0079] The present invention provides Figure 1 The real-time dynamic optimization and adjustment method for the tunneling process shown includes the following steps:

[0080] Acquire real-time operating status data of tunneling equipment, including propulsion force, cutterhead speed, support pressure, and formation feedback parameters, while recording the dynamic change characteristics of the formation;

[0081] The data collection step further includes multi-sensor data fusion technology to synchronously process tunneling parameters from different types of sensors to improve data accuracy and reliability;

[0082] Sensors include, but are not limited to, laser range sensors, inertial measurement units (IMUs), formation pressure sensors, and accelerometers;

[0083] All data is first filtered and denoised after collection. The Kalman filter algorithm is used to reduce data fluctuations, and a time synchronization mechanism is used to ensure that data with different sampling frequencies can be compared and analyzed in the same time dimension.

[0084] In addition, in order to further improve the accuracy of stratum characteristic identification, the current excavation parameters are dynamically calibrated by combining historical excavation data and stratum geological exploration information to reduce data anomalies caused by stratum mutations, thereby ensuring that the dynamic optimization and adjustment of excavation parameters can be carried out on the basis of reliable data information.

[0085] Based on the collected data, a nonlinear dynamic model of the tunneling process is constructed, and adaptive algorithms are used to identify the response characteristics of the formation and establish a mathematical relationship between the interaction between the equipment and the formation.

[0086] The nonlinear dynamic modeling step uses a deep neural network (DNN) combined with finite element analysis (FEA) to accurately model the interaction between the roadheader and the formation. The specific steps are as follows:

[0087] First, a physical model of the ground-roadhead interaction was constructed using the finite element analysis method to simulate the effects of different thrusts and cutterhead torques on ground disturbance.

[0088] Subsequently, a deep neural network was used for data training, enabling the model to learn the complex relationship between tunneling parameters and formation feedback from historical data and achieve adaptive adjustments.

[0089] In addition, the genetic algorithm (GA) is further used to optimize the neural network parameters, so that the model can respond quickly when the formation changes and accurately predict the resonance critical value under different tunneling conditions.

[0090] Using time series analysis combined with ground stiffness and roadheader vibration characteristics, the resonance threshold under the current tunneling state is calculated to determine whether there is a low-frequency resonance trend and to provide an early warning of potential instability risks.

[0091] The resonance risk prediction step is not only based on traditional time series analysis methods, but also incorporates anomaly detection technology based on machine learning to more efficiently identify low-frequency resonances that occur during tunneling.

[0092] The vibration spectrum of the roadheader was analyzed using the autoregressive integrated moving average (ARIMA) model, and the vibration characteristics were classified using the support vector machine (SVM) classification model to predict the resonance mode.

[0093] When abnormal resonance trends are detected, the tunneling strategy is automatically adjusted, such as reducing thrust, adjusting cutterhead speed, or optimizing support pressure, to avoid entering the resonance zone.

[0094] In addition, the prediction process uses the Bayesian update mechanism to continuously optimize the model parameters, so that the prediction accuracy continues to improve during the long-term operation.

[0095] The resonance risk prediction step uses the generalized resonance calculation model (GRCM) to calculate the resonance critical value through the following steps;

[0096] Calculate the main resonance frequency of the roadheader using the following expression:

[0097]

[0098] , where fr is the natural resonance frequency of the roadheader, that is, the natural frequency of the free vibration of the system without external disturbance, k eq is the equivalent stiffness of the tunnel boring machine and the stratum, which represents the elastic restoring force coefficient when the stratum contacts the tunnel boring machine. This value is affected by factors such as stratum type, support pressure, and cutterhead design. m eq is the equivalent mass, which represents the inertial contribution to the vibration of the roadheader structure and rotating parts, including the cutterhead, propulsion system and structural support parts;

[0099] Calculate the damping characteristics of the formation, the calculation expression is as follows:

[0100]

[0101] , where ξ is the damping ratio, which reflects the ability of the formation to attenuate vibration, c eq is the equivalent damping coefficient, which describes the energy dissipation capacity of the ground material to the vibration of the tunnel boring machine;

[0102] Calculate the resonance gain factor of the system. The calculation expression is as follows:

[0103]

[0104] , where G r is the resonance gain factor, which indicates the degree of coupling between the current tunneling frequency and the natural frequency of the system, and f is the actual vibration frequency of the tunneling machine, which is measured in real time by the sensor;

[0105] To judge the resonance risk, the calculation expression is as follows:

[0106] G r >G th

[0107] , where G th is the resonance gain factor reference threshold;

[0108] If the resonance gain factor G r Exceeding the set threshold G th , a resonance warning is triggered and the tunneling parameters are adjusted to avoid the resonance risk.

[0109] Based on the prediction results, key tunneling parameters such as propulsion force, cutterhead speed, and support pressure are dynamically adjusted. Adaptive control algorithms are used to optimize the tunneling mode, allowing the equipment to operate outside the resonance range while ensuring tunneling efficiency.

[0110] The specific steps for real-time optimization and adjustment are as follows:

[0111] Adopting multi-objective optimization control strategy to improve excavation efficiency while ensuring excavation stability;

[0112] The optimization strategy uses the particle swarm optimization algorithm (PSO) to establish a nonlinear mapping relationship between multiple key parameters such as propulsion force, cutterhead speed, and support pressure, and find the optimal adjustment solution in the shortest time.

[0113] In addition, the optimization strategy is combined with fuzzy control theory. When the ground conditions suddenly change, the operating parameters of the tunnel boring machine can be quickly adjusted, and through real-time feedback control, the tunneling process can always be in an efficient and safe working state.

[0114] The real-time optimization and adjustment step uses a nonlinear time-varying optimization algorithm (NTOA) to calculate the optimal tunneling parameter adjustment values. The specific steps are as follows:

[0115] Construct a nonlinear constrained optimization problem, and the calculation expression is as follows:

[0116]

[0117] , where J is the objective function, P i is the current value of the ith tunneling parameter, is the optimal value of the i-th excavation parameter, that is, the parameter value determined after optimization calculation, w i is the weight factor, which indicates the importance of the i-th parameter to the objective function, and N is the total number of tunneling parameters that need to be optimized;

[0118] Define the constraints and the calculation expression is as follows:

[0119] P min ≤P i ≤P max

[0120] , where P max and P max are the minimum and maximum values of the tunneling parameters, respectively;

[0121] The Lagrange multiplier method is used to solve the problem. The calculation expression is as follows:

[0122]

[0123] , where is the Lagrangian function, which combines the optimization objective with the constraints to solve the optimal parameters. λ is the Lagrangian multiplier, which is used to adjust the constraint strength of the optimization problem.

[0124] Iteratively solve the optimal excavation parameters, and the calculation expression is as follows:

[0125]

[0126] , where is the updated excavation parameter. After optimization and adjustment, it ensures that the excavation process avoids the resonance range and improves efficiency. α is the step size factor, which controls the rate of optimization adjustment. It is usually obtained through experience or adaptive adjustment. It is the partial derivative of the objective function, which indicates the direction of influence of the current parameters on the optimization target;

[0127] Adjust the tunnel boring machine parameters in real time to ensure the equipment operates in the optimal state and avoid resonance risks.

[0128] Real-time monitoring of the impact of adjusted tunneling parameters on ground stability and equipment load. If resonance is still caused by parameter adjustments, further optimization of tunneling strategies is performed to form a closed-loop control system.

[0129] The specific steps of dynamic feedback control are as follows:

[0130] Adaptive neuro-fuzzy inference system (ANFIS) is used for control optimization, which can adjust the control rules according to the changes in the formation and the status of the equipment during the tunneling process;

[0131] First, the historical excavation data is classified through the neural network self-learning mechanism, and the fuzzy control rules are adjusted according to the current excavation status;

[0132] Then, fuzzy logic reasoning is used to evaluate the operating status of the tunnel boring machine in real time to determine whether the current tunneling parameters are in the resonance risk range;

[0133] Finally, through fuzzy weight adjustment, the tunneling parameter adjustment strategy is continuously optimized to ensure the stability and efficiency of the tunneling system.

[0134] When unpredictable ground changes or abnormal equipment vibrations are detected, the emergency mechanism is automatically triggered to adjust the excavation rhythm, reduce the cutterhead load, and optimize the adjustment strategy based on historical data.

[0135] Implementation Method 1: During shield machine or tunnel boring machine (TBM) construction, tunneling parameter adjustments must closely match the changing characteristics of the stratum to prevent resonance and ensure stable equipment operation. This implementation method proposes an adaptive optimization method for tunneling parameters based on intelligent feedback control. By leveraging multi-sensor data fusion, nonlinear dynamic modeling, and fuzzy control algorithms, this method dynamically adjusts tunneling parameters, thereby improving construction efficiency, reducing equipment losses, and effectively avoiding low-frequency resonance caused by delayed parameter adjustment.

[0136] First, this implementation utilizes multi-sensor data fusion technology to collect real-time data on thrust, cutterhead speed, support pressure, equipment vibration frequency, and formation feedback parameters involved in the tunneling process. During construction, shield machines or TBMs are significantly affected by formation conditions, which can exhibit characteristics such as heterogeneity and mutation. Therefore, data from a single sensor often struggles to fully reflect the actual working conditions and is susceptible to noise interference, which can even lead to misjudgments. To address this issue, this method utilizes an inertial measurement unit (IMU), laser ranging sensors, formation pressure sensors, accelerometers, and other high-precision sensing equipment to perform multi-dimensional sensing of tunneling data, and uses data fusion technology to improve the accuracy and reliability of the data.

[0137] The collected data is first denoised using a Kalman filter algorithm to remove high-frequency noise and environmental interference. Subsequently, a Bayesian update mechanism is used to dynamically weight and optimize the sensor data, improving the reliability of real-time data under abnormal conditions. For example, when a shield machine is tunneling in complex formations (such as sand layers or fault zones), individual sensors may experience deviations due to external interference. In this case, the Bayesian update mechanism can combine historical data with other sensor information to correct abnormal data, thereby improving the stability and reliability of the overall data.

[0138] After obtaining high-precision real-time data, the intelligent control system uses nonlinear dynamic modeling methods to construct a mathematical model of the shield machine-stratum interaction to predict the resonance risk during the excavation process. Since the relationship between the tunnel boring machine and the stratum is highly nonlinear, and the effects of propulsion force, cutterhead speed, and support pressure on vibration response vary greatly under different stratum conditions, linear modeling cannot accurately characterize the dynamic characteristics of the excavation process. To this end, this embodiment uses finite element analysis (FEA) combined with deep neural network (DNN) methods to simulate and calculate the stress conditions and resonance trends of the shield machine under different excavation states.

[0139] Specifically, finite element analysis is used to establish a physical and mechanical model of the shield machine and the formation, including the effects of the thrust cylinder force, cutterhead shear force, and support pressure on formation disturbances, and to analyze its vibration characteristics under different working conditions. At the same time, based on historical excavation data, the deep neural network training model can learn the system response under different excavation parameter combinations from a large amount of historical data, thereby adaptively identifying resonance risks during the excavation process and predicting possible instability in advance. In addition, this modeling method also combines the support vector machine (SVM) classification algorithm to classify the excavation status, ensuring that the system can accurately identify possible vibration modes under different formation conditions and adjust the excavation parameters to avoid entering the resonance range.

[0140] Once the system identifies the risk of resonance, the intelligent control system dynamically adjusts the thrust, cutterhead speed, and support pressure using a fuzzy control algorithm to ensure that the tunneling process remains within a stable range. Fuzzy control is an intelligent control method based on expert experience rules, suitable for solving the complex adjustment problems caused by the uncertainty of formation parameters during shield machine tunneling. This method uses multiple fuzzy rules, such as:

[0141] If the vibration frequency is close to the resonance range and the formation is soft, reduce the propulsion force and cutterhead speed;

[0142] If the vibration frequency is within the stable range but the formation pressure fluctuates greatly, the support pressure should be appropriately increased to improve soil stability;

[0143] If the system detects a sudden change in the stratum (such as entering an area with high water content), the excavation speed will be reduced and the grouting parameters will be optimized to improve the stratum support capacity.

[0144] The fuzzy control system dynamically optimizes these rules and, in conjunction with continuously updated ground data during tunneling, makes tunneling parameter adjustments more adaptive and real-time. Furthermore, the system incorporates a closed-loop feedback mechanism. After adjusting tunneling parameters, the system continues to monitor the equipment's response and continuously optimizes its adjustment strategy, ensuring the tunneling system remains in optimal working condition.

[0145] In addition to optimizing tunneling parameters, this method also provides equipment health management capabilities to prevent damage to critical roadheader components caused by long-term resonance. For example, the operating conditions of key components such as the thrust cylinder, main bearing, and cutterhead are affected by tunneling vibrations. Long-term exposure to high vibrations can accelerate fatigue damage and reduce equipment lifespan. To mitigate this issue, this implementation utilizes a machine learning algorithm to establish an equipment health monitoring model, analyzing the stresses on key components in real time and predicting potential damage trends.

[0146] When the system detects a component experiencing prolonged high stress, it issues an early warning, prompting the operator to perform maintenance or automatically adjusts tunneling parameters to reduce the impact load on the component. Furthermore, the system uses reinforcement learning algorithms to continuously optimize tunneling parameter adjustment strategies over the long-term tunneling process, ensuring the equipment maintains optimal operating conditions under varying conditions, thereby extending the life of the tunneling machine and improving construction safety.

[0147] The application of this implementation method can effectively improve the operating efficiency of the tunnel boring machine while reducing equipment wear and maintenance costs. In actual engineering applications, this method can be widely applied to complex tunneling projects such as urban subway shield construction, mountain tunnel excavation, and mine roadway excavation. It is particularly suitable for construction environments with complex strata and high vibration risks. By utilizing technologies such as multi-sensor data fusion, intelligent feedback control, and fuzzy control optimization, this method can achieve real-time dynamic adjustment of tunneling parameters during the tunneling process, avoiding resonance caused by adjustment lags and ensuring stable and efficient operation of tunneling equipment under various complex working conditions.

[0148] In summary, this implementation utilizes adaptive optimization of tunneling parameters based on intelligent feedback control. By leveraging key technologies such as multi-sensor data fusion, high-precision modeling, fuzzy control optimization, and equipment health management, it enables efficient and stable tunneling of shield machines or TBMs in complex strata. This approach not only improves construction safety and reduces equipment maintenance costs, but also continuously optimizes tunneling parameter adjustment strategies over the long term, providing an efficient and intelligent solution for tunneling, mining, and other tunneling projects.

[0149] Implementation Method 2: During the construction of a roadheader or tunnel boring machine (TBM), due to the complexity of the ground conditions and the uncertainty of the construction environment, the occurrence of resonance phenomena is often difficult to accurately predict. Sudden vibration anomalies may even occur during the construction process, affecting the excavation efficiency and increasing equipment loss and safety risks. This implementation method proposes a resonance risk prediction and optimization adjustment method based on machine learning. It combines advanced algorithms such as finite element analysis (FEA), deep neural network (DNN), support vector machine (SVM), and particle swarm optimization (PSO) to achieve high-precision prediction of resonance risks during the excavation process. Based on the prediction results, the excavation parameters are intelligently optimized and adjusted to ensure that the excavation system can maintain a stable and efficient operation state under different ground conditions.

[0150] In order to improve the prediction accuracy of the tunneling process, this embodiment first uses the finite element analysis (FEA) method to establish a high-precision physical model of the interaction between the tunnel boring machine and the formation. During the tunneling process, the cutterhead, propulsion cylinder, support system and other components of the shield machine or TBM will exert forces on the formation, and factors such as the structure, soil quality, and moisture content of the formation determine the vibration response of the equipment. Therefore, the dynamic relationship between the tunneling equipment and the formation is a highly complex nonlinear problem that cannot be accurately described by a simple empirical model. Based on this, this embodiment uses finite element analysis to simulate and analyze the operating status of the tunnel boring machine under different formation conditions, establish a coupled dynamic model of the equipment and formation, and calculate the force conditions and possible vibration responses of the formation under different tunneling parameters (propulsion force, cutterhead speed, support pressure, etc.).

[0151] During the physical modeling process, the material parameters of the stratum were first modeled, including key mechanical parameters such as density, elastic modulus, Poisson's ratio, cohesion, and internal friction angle. Subsequently, the mechanical structure of the tunneling equipment was modeled to analyze the stress distribution of various components under load. Finally, using finite element analysis, the stratum deformation and equipment vibration characteristics under different tunneling parameters were calculated, forming a high-precision dynamics database that provided reliable data support for subsequent machine learning model training.

[0152] After obtaining a highly accurate physical model, a deep neural network (DNN) was further used for data learning and feature extraction to achieve real-time prediction of resonance risk during tunneling. Because the operating state of the tunneling system is affected by a variety of complex factors, such as sudden changes in the ground, thrust adjustment, cutterhead load fluctuations, and pressure changes in the support system, traditional linear prediction models struggle to accurately describe the interactions between these factors. Therefore, a deep neural network can be used to learn the nonlinear relationship between tunneling parameters and resonance risk from large-scale historical data, thereby improving the accuracy and real-time nature of predictions.

[0153] The DNN model training process includes key steps such as data preprocessing, feature extraction, model training, and optimization. First, historical excavation data is standardized to remove outliers, and principal component analysis (PCA) is used for dimensionality reduction to reduce data redundancy and improve computational efficiency. Subsequently, the DNN is used to extract deep features between excavation parameters and ground vibrations. Combined with a long short-term memory network (LSTM), the time series changes in parameters during excavation are modeled to more accurately predict possible resonance phenomena in the future. After training, the DNN model can analyze real-time data during actual excavation and output a resonance risk index (RRI) to determine whether there is a resonance risk in the current excavation state.

[0154] (3) Resonance pattern recognition based on support vector machine

[0155] On the basis of DNN prediction, this embodiment further introduces the support vector machine (SVM) classification algorithm to classify the resonance modes that may appear during the excavation process, so that the system can adopt more accurate optimization and adjustment strategies in different situations. SVM is a classification method based on statistical learning theory, which is suitable for small sample and high-dimensional data analysis and has high classification accuracy in resonance identification problems. During the training process of the SVM model, multiple resonance categories are first extracted from the historical excavation data, such as cutterhead torque resonance, thrust cylinder resonance, formation-equipment coupling resonance, etc., and the characteristics of each resonance mode are labeled. Subsequently, the SVM classifier is used to perform pattern recognition on the real-time data, and the optimal excavation parameter adjustment strategy is selected based on the recognition results.

[0156] When the system detects that tunneling parameters enter a known resonance mode, the SVM classifier automatically matches the corresponding optimization adjustment plan. For example, if the propulsion cylinder resonance is detected, the system may reduce the propulsion force and adjust the cutterhead speed to reduce system impact. If the formation-equipment coupling resonance is detected, the system may adjust the support pressure to change the formation stress state and avoid further resonance amplification.

[0157] After completing resonance risk prediction and pattern recognition, this implementation utilizes a particle swarm optimization (PSO) algorithm to optimally adjust tunneling parameters to ensure stable system operation and improve construction efficiency while ensuring safety. PSO is an optimization algorithm based on swarm intelligence that simulates the collaborative behavior of biological populations to quickly find the global optimal solution. During tunneling parameter optimization, the PSO algorithm aims to find the optimal combination of multiple variables, such as propulsion force, cutterhead speed, and support pressure, to maximize tunneling efficiency while avoiding resonance.

[0158] The optimization process includes the following steps:

[0159] Initialize particle swarm: set a set of candidate tunneling parameter combinations, where each particle represents a possible parameter setting.

[0160] Calculate the fitness function: Using the resonance risk index (RRI) predicted by physical modeling and machine learning, define the fitness function: F = w1·tunneling efficiency-w2·resonance risk index, where w1 and w2 are weight factors used to balance efficiency and safety.

[0161] Update particle velocity and position: Dynamically adjust tunneling parameters based on the deviation between the historical optimal solution and the current solution.

[0162] Iterative optimization: Continuously search for better parameter combinations until the optimization goal is met.

[0163] Ultimately, the PSO algorithm will provide the optimal excavation parameter combination, which will be adjusted in real time by the intelligent control system to ensure that the equipment operates efficiently while avoiding entering the resonance range.

[0164] The application of this implementation effectively enhances the intelligence of the tunneling process, ensuring stable and efficient operation of the tunnel boring machine under diverse ground conditions. By combining physical modeling, machine learning prediction, resonance pattern recognition, and intelligent optimization, this method can predict resonance risks before construction begins and implement dynamic optimization adjustments during the tunneling process, thereby reducing equipment damage caused by resonance, lowering maintenance costs, and improving overall construction efficiency.

[0165] This method is applicable to complex engineering environments such as subway tunnel construction, mine tunneling, and large-scale infrastructure development. It is particularly well-suited for construction areas with variable ground conditions and high resonance risk. Through machine learning and intelligent optimization techniques, this method continuously learns and optimizes tunneling parameter adjustment strategies over the long term, ultimately achieving a smarter and safer tunneling control system.

[0166] Implementation method three: During the construction of the tunnel boring machine, the stratum environment usually has nonlinear, time-varying and sudden characteristics, making the adjustment of tunneling parameters an extremely challenging task. The traditional fixed parameter control strategy cannot adapt to complex stratum changes, which can easily cause the tunnel boring machine to enter the resonance range, thereby causing equipment loss, reduced tunneling efficiency and even construction accidents. This implementation method proposes a dynamic tunneling control method based on the nonlinear time-varying optimization algorithm (NTOA), using the generalized resonance calculation model (GRCM), autoregressive integral moving average (ARIMA) model, Lagrangian optimization method and closed-loop feedback control and other technologies to achieve accurate calculation and dynamic adjustment of tunneling parameters, thereby improving the stability and construction efficiency of the tunnel boring machine.

[0167] During the excavation process, there is a complex nonlinear relationship between the operating state of the equipment and the characteristics of the stratum, and the characteristics of the stratum will continue to change as the excavation goes deeper, showing time-varying characteristics. If the excavation system uses a fixed parameter control strategy, when the excavation enters a new stratum area, the original parameters may no longer be applicable, resulting in abnormal operation of the tunnel boring machine. For example, when the tunnel boring machine enters a high-water-content clay layer from a dense sand layer, if the original propulsion force and cutterhead speed are still maintained, it may cause cutterhead torque overload, propulsion obstruction, and even cause equipment resonance. Therefore, this embodiment proposes a method based on a nonlinear time-varying optimization algorithm (NTOA) to calculate the optimal excavation parameters in real time and dynamically adjust the excavation strategy to adapt to complex stratum changes and improve construction stability and efficiency.

[0168] The core concept of NTOA is to utilize historical data and real-time monitoring information to construct a dynamic optimization model, enabling tunneling parameters to be adjusted adaptively to current ground conditions. This approach first models the tunneling process using a data-driven approach, incorporating predictive algorithms to proactively identify potential resonance risks. It then utilizes nonlinear optimization methods to determine the optimal combination of thrust, cutterhead speed, and support pressure, ensuring the tunneling system maintains efficient and stable operation.

[0169] During tunneling, resonance often results from delayed adjustment of tunneling parameters or dynamic instability caused by sudden changes in the ground. Therefore, it's crucial to identify potential risks and optimize adjustments before the tunneling machine enters the resonance range. This implementation utilizes a generalized resonance calculation model (GRCM), comprehensively considering the tunneling equipment's stiffness and damping characteristics, ground heterogeneity, and vibration modes. This model calculates the resonance threshold under the current tunneling state and analyzes the tunneling system's resonance trends using real-time data.

[0170] To improve prediction accuracy, this method further employs an autoregressive integrated moving average (ARIMA) model to perform time series analysis on vibration data collected during tunneling. The ARIMA model leverages the changing trends of historical data to predict the vibration patterns of the tunneling system over a period of time and determine whether current tunneling parameters are likely to cause resonance. If the prediction indicates that the tunneling system may enter a resonance state, the optimization control system proactively adjusts the tunneling parameters to avoid the resonance zone, ensuring stable operation of the equipment.

[0171] After identifying resonance risk, the system needs to quickly calculate the optimal tunneling parameter adjustment strategy to maximize tunneling efficiency while ensuring construction safety. To this end, this implementation uses a Lagrangian optimization method to transform the tunneling parameter optimization problem into a nonlinear constrained optimization problem, solving it with the goals of maximizing tunneling efficiency and minimizing resonance risk.

[0172] During the optimization process, the system first determines the stress characteristics of the current formation and evaluates the operating status of the TBM. It then constructs an objective function, using tunneling speed, cutterhead speed, thrust, and support pressure as optimization variables, and seeks the optimal solution within the constraints. For example, if the TBM is in an area of high vibration risk, the optimization algorithm might automatically reduce cutterhead speed and increase thrust to minimize vibration. It also adjusts support pressure to achieve more uniform ground stress, thereby reducing the risk of resonance in the tunneling system.

[0173] The key advantage of this optimization method lies in its ability to rapidly determine optimal parameters within complex nonlinear systems and adapt to time-varying stratum characteristics. Compared to traditional fixed-parameter control methods, this method dynamically adjusts tunneling parameters, ensuring the tunnel boring machine consistently operates at optimal levels in varying strata, improving tunneling efficiency and reducing equipment wear.

[0174] After the optimization calculations are complete, the system uses closed-loop feedback control to adjust the TBM's operating status in real time, ensuring that tunneling parameters remain within the optimal range. The basic principle of closed-loop feedback control is that the system monitors the TBM's operating status in real time and compares it with the optimal parameters calculated by the optimization algorithm. If the current tunneling parameters deviate from the optimal values, the system automatically makes fine adjustments to ensure tunneling stability.

[0175] For example, when the TBM is passing through soft ground, if the system detects that the cutterhead speed is too high, causing increased vibration, it will automatically reduce the cutterhead speed and increase the thrust accordingly to compensate for torque loss. Simultaneously, the support pressure is adjusted accordingly to ensure ground stability. This real-time adjustment mechanism ensures the TBM maintains optimal operating conditions under varying operating conditions, avoiding resonance caused by improper parameter settings, thereby improving construction safety and tunneling efficiency.

[0176] The system also features self-learning capabilities, enabling it to continuously optimize control strategies based on data accumulated over long periods of tunneling. For example, as the tunnel boring machine traverses different strata, the system records the optimal tunneling parameters for each stratum and stores these data in a historical experience database. When encountering similar strata in the future, the system can directly draw on these historical experiences, improving the response time of tunneling parameter adjustments, reducing calculation time, and ultimately increasing construction efficiency.

[0177] The application of this implementation effectively enhances the intelligence level of the tunneling system, enabling precise parameter adjustment of the roadheader in complex and variable ground conditions, reducing safety risks caused by resonance while improving tunneling efficiency. By combining advanced technologies such as nonlinear time-varying optimization algorithms, generalized resonance computational models, time series analysis, Lagrangian optimization methods, and closed-loop feedback control, this method ensures that the roadheader maintains efficient and stable operation over long-term construction periods.

[0178] This method is applicable to complex tunneling projects such as urban subway construction, mountain tunneling, and mine roadway excavation. It is particularly well-suited for tunneling scenarios characterized by frequent ground changes, complex construction environments, and high resonance risks. By monitoring the tunneling system's operating status in real time and dynamically adjusting tunneling parameters using advanced optimization algorithms, this method not only improves construction efficiency and reduces equipment wear, but also continuously optimizes tunneling strategies over the long term, providing a safer and more intelligent control solution for tunneling projects.

[0179] The present invention optimizes and adjusts the tunneling parameters in real time, combines nonlinear dynamic modeling and machine learning prediction, and effectively prevents the tunnel boring machine from entering the resonance range during the construction process. Traditional tunneling control methods mainly rely on manual experience or fixed parameter adjustment, and are unable to respond to the dynamic changes of the stratum in real time, resulting in a lag in the adjustment of the propulsion force, cutterhead speed and support pressure, which may trigger low-frequency resonance, causing fatigue damage to the cutterhead, uneven force on the propulsion cylinder, and even equipment instability. The present invention uses a generalized resonance calculation model (GRCM) and a nonlinear time-varying optimization algorithm (NTOA) to calculate the resonance critical value in real time, and uses closed-loop feedback control for dynamic adjustment to ensure that the tunneling parameters are always within a safe range. This method can significantly reduce the risk of damage to mechanical components caused by resonance, improve the long-term reliability of the tunnel boring machine, while reducing equipment maintenance costs and downtime, thereby improving construction efficiency.

[0180] The present invention adopts a multi-objective optimization control strategy to improve the excavation efficiency while ensuring the stable operation of the equipment. Traditional excavation methods usually adopt a fixed parameter operation mode. When encountering changes in the formation, manual intervention is often required to adjust the excavation parameters, resulting in limited construction efficiency, and even problems such as propulsion obstruction or cutterhead jamming when sudden changes in formation conditions occur. The present invention uses a particle swarm optimization (PSO) algorithm and a deep neural network (DNN) prediction model to automatically calculate the optimal combination of propulsion force, cutterhead speed and support pressure based on formation feedback data, and dynamically adjust the excavation mode to avoid the equipment running in an unnecessary inefficient state. In addition, the method can learn the optimal parameter control strategy in the excavation process in real time through an adaptive neural fuzzy inference system (ANFIS), so that the system has self-learning and self-optimization capabilities, reduces the need for manual intervention, and improves the level of construction automation, thereby maximizing the excavation speed while ensuring construction safety.

[0181] The present invention adopts a stratum prediction and optimization adjustment method based on machine learning to effectively improve the adaptability of the tunnel boring machine under different stratum conditions. When encountering sudden changes in strata (such as soft and hard alternating layers, fault fracture zones, water-rich sand layers, etc.), traditional tunneling methods are often unable to respond quickly, resulting in untimely adjustment of tunneling parameters, which may cause equipment resonance, stratum instability, shield posture deviation and other problems, seriously affecting construction safety. The present invention uses a support vector machine (SVM) classification algorithm and an autoregressive integrated moving average (ARIMA) model to monitor the changes in the mechanical properties of the stratum during tunneling in real time, predict the construction risks that may be brought about by sudden changes in strata in advance, and automatically adjust the tunneling parameters to ensure that the construction process is stable and controllable. In addition, the present invention adopts an abnormal state emergency response mechanism. When unpredictable stratum mutations or abnormal vibrations of equipment are detected, the system can automatically adjust the tunneling rhythm, reduce the cutterhead load, and optimize the adjustment strategy in combination with historical data, thereby minimizing construction safety hazards and improving the reliability of the tunneling system in complex stratum environments.

[0182] The present invention provides Figure 2 The real-time dynamic optimization and adjustment system for the tunneling process shown in the figure includes a data acquisition and status monitoring module, a nonlinear dynamic modeling module, a resonance risk prediction module, an adaptive optimization and adjustment module, a closed-loop feedback control module, and an abnormal state emergency processing module:

[0183] The data acquisition and status monitoring module acquires real-time operating status data of the tunneling equipment, including propulsion force, cutterhead speed, support pressure, and formation feedback parameters, while also recording the dynamic characteristics of the formation.

[0184] The nonlinear dynamic modeling module builds a nonlinear dynamic model of the tunneling process based on the collected data, uses an adaptive algorithm to identify the response characteristics of the formation, and establishes a mathematical relationship between the interaction between the equipment and the formation;

[0185] The resonance risk prediction module uses a time series analysis method combined with the ground stiffness and vibration characteristics of the tunnel boring machine to calculate the resonance critical value under the current tunneling state, determine whether there is a low-frequency resonance trend, and warn of potential instability risks;

[0186] The adaptive optimization and adjustment module dynamically adjusts key tunneling parameters such as propulsion force, cutterhead speed, and support pressure based on prediction results. It optimizes the tunneling mode through an adaptive control algorithm, allowing the equipment to operate outside the resonance range while ensuring tunneling efficiency.

[0187] A closed-loop feedback control module monitors in real time the impact of adjusted tunneling parameters on ground stability and equipment load. If the parameter adjustment still causes resonance, the tunneling strategy is further optimized to form a closed-loop control.

[0188] The abnormal state emergency processing module automatically triggers the emergency mechanism when unpredictable formation mutations or abnormal equipment vibrations are detected, adjusts the excavation rhythm, reduces the cutterhead load, and optimizes the adjustment strategy based on historical data.

[0189] The real-time dynamic optimization and adjustment method for the excavation process provided in an embodiment of the present invention is realized by the above-mentioned real-time dynamic optimization and adjustment system for the excavation process. The specific methods and processes of the real-time dynamic optimization and adjustment system for the excavation process are detailed in the embodiment of the real-time dynamic optimization and adjustment method for the excavation process, and will not be repeated here.

[0190] The above formulas are all dimensionless and numerical calculations. The formulas are obtained by collecting a large amount of data and performing software simulation to obtain the most recent real situation. The preset parameters in the formulas are set by technicians in this field according to actual conditions.

[0191] The above description is merely illustrative of certain exemplary embodiments of the present invention. It goes without saying that those skilled in the art will be able to modify the described embodiments in various ways without departing from the spirit and scope of the present invention. Therefore, the above drawings and description are illustrative in nature and should not be construed as limiting the scope of protection of the claims.

[0192] It should be noted that, in this document, if there are relational terms such as first and second, etc., they are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "comprises", "comprising" or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device that includes a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device. In the absence of further restrictions, an element defined by the sentence "comprising a ..." does not exclude the presence of other identical elements in the process, method, article or device that includes the element.

[0193] It should be understood that in the various embodiments of the present application, the size of the serial numbers of the above-mentioned processes does not mean the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present application.

[0194] Those skilled in the art will appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0195] Those skilled in the art will clearly understand that, for the convenience and brevity of description, the specific working processes of the systems, devices and units described above can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0196] The units described as separate components may or may not be physically separate, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed across multiple network units. Some or all of these units may be selected to achieve the purpose of this embodiment according to actual needs.

[0197] In addition, each functional unit in each embodiment of the present application may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit.

[0198] The above description is merely a specific embodiment of the present application, but the scope of protection of the present application is not limited thereto. Any changes or substitutions that can be easily conceived by a person skilled in the art within the technical scope disclosed in this application should be included in the scope of protection of this application. Therefore, the scope of protection of this application should be based on the scope of protection of the claims.

[0199] The above description is merely illustrative of certain exemplary embodiments of the present invention. It goes without saying that those skilled in the art will be able to modify the described embodiments in various ways without departing from the spirit and scope of the present invention. Therefore, the above drawings and description are illustrative in nature and should not be construed as limiting the scope of protection of the claims.

Claims

1. A real-time dynamic optimization and adjustment method for the tunneling process, characterized in that: The following steps are involved: Acquire the operating status data of tunneling equipment in real time and record the dynamic change characteristics of the stratum; Based on the collected data, a nonlinear dynamic model of the tunneling process is constructed, and adaptive algorithms are used to identify the response characteristics of the formation and establish a mathematical relationship between the interaction between the equipment and the formation. Using time series analysis combined with ground stiffness and roadheader vibration characteristics, the resonance threshold under the current tunneling state is calculated to determine whether there is a low-frequency resonance trend and to provide an early warning of potential instability risks. Based on the prediction results, key tunneling parameters are dynamically adjusted, and the tunneling mode is optimized through adaptive control algorithms to avoid the resonance range of the equipment while ensuring tunneling efficiency. Real-time monitoring of the impact of adjusted tunneling parameters on ground stability and equipment load. If resonance is still caused by parameter adjustments, further optimization of tunneling strategies is performed to form a closed-loop control system. When unpredictable ground changes or abnormal equipment vibrations are detected, the emergency mechanism is automatically triggered to adjust the excavation rhythm, reduce the cutterhead load, and optimize the adjustment strategy based on historical data.

2. The real-time dynamic optimization and adjustment method for the tunneling process according to claim 1 is characterized in that: The data collection step further includes multi-sensor data fusion technology to synchronously process tunneling parameters from different types of sensors to improve data accuracy and reliability; Sensors include, but are not limited to, laser range sensors, inertial measurement units, formation pressure sensors, and accelerometers; All data is first filtered and denoised after collection. The Kalman filter algorithm is used to reduce data fluctuations, and a time synchronization mechanism is used to ensure that data with different sampling frequencies can be compared and analyzed in the same time dimension. In addition, in order to further improve the accuracy of stratum characteristic identification, the current excavation parameters are dynamically calibrated by combining historical excavation data and stratum geological exploration information to reduce data anomalies caused by stratum mutations, thereby ensuring that the dynamic optimization and adjustment of excavation parameters can be carried out on the basis of reliable data information.

3. The real-time dynamic optimization and adjustment method for the tunneling process according to claim 1 is characterized in that: The nonlinear dynamic modeling step uses a deep neural network combined with finite element analysis to accurately model the interaction between the roadheader and the formation. The specific steps are as follows: First, a physical model of the ground-roadhead interaction was constructed using the finite element analysis method to simulate the effects of different thrusts and cutterhead torques on ground disturbance. Subsequently, a deep neural network was used for data training, enabling the model to learn the complex relationship between tunneling parameters and formation feedback from historical data and achieve adaptive adjustments. In addition, genetic algorithms were further used to optimize the neural network parameters, so that the model can respond quickly when the formation changes and accurately predict the resonance critical value under different tunneling conditions.

4. The real-time dynamic optimization and adjustment method for the tunneling process according to claim 1 is characterized in that: The resonance risk prediction step is not only based on traditional time series analysis methods, but also incorporates anomaly detection technology based on machine learning to more efficiently identify low-frequency resonances that occur during tunneling. The vibration spectrum of the roadheader is analyzed in time series using an autoregressive integrated moving average model. The vibration characteristics are then classified using a support vector machine classification model to predict the resonance mode. Automatically adjust the tunneling strategy when abnormal resonance trends are detected; In addition, the prediction process uses the Bayesian update mechanism to continuously optimize the model parameters, so that the prediction accuracy continues to improve during the long-term operation.

5. The real-time dynamic optimization and adjustment method for the tunneling process according to claim 1 is characterized in that: The resonance risk prediction step uses a generalized resonance calculation model to calculate the resonance critical value through the following steps: Calculate the main resonance frequency of the roadheader using the following expression: ， Where, f r is the natural resonance frequency of the roadheader, k eq is the equivalent stiffness of the tunnel boring machine and the ground, m eq is the equivalent mass; Calculate the damping characteristics of the formation, the calculation expression is as follows: ， Where ξ is the damping ratio, c eq is the equivalent damping coefficient; Calculate the resonance gain factor of the system. The calculation expression is as follows: ， Where G r is the resonance gain factor, f is the actual vibration frequency of the roadheader; To judge the resonance risk, the calculation expression is as follows: G r >G th , Where G th is the resonance gain factor reference threshold; If the resonance gain factor G r Exceeding the set threshold G th , a resonance warning is triggered and the tunneling parameters are adjusted to avoid the resonance risk.

6. The real-time dynamic optimization and adjustment method for the tunneling process according to claim 1, characterized in that: The specific steps for real-time optimization and adjustment are as follows: Adopting multi-objective optimization control strategy to improve excavation efficiency while ensuring excavation stability; The optimization strategy uses the particle swarm optimization algorithm to establish a nonlinear mapping relationship between key parameters and find the optimal adjustment solution in the shortest time; In addition, the optimization strategy is combined with fuzzy control theory. When the ground conditions suddenly change, the operating parameters of the tunnel boring machine can be quickly adjusted, and through real-time feedback control, the tunneling process can always be in an efficient and safe working state.

7. The real-time dynamic optimization and adjustment method for the tunneling process according to claim 1, characterized in that: The real-time optimization and adjustment step uses a nonlinear time-varying optimization algorithm to calculate the optimal tunneling parameter adjustment value. The specific steps are as follows: Construct a nonlinear constrained optimization problem, and the calculation expression is as follows: ， Where J is the objective function, P i is the current value of the ith tunneling parameter, is the optimal value of the i-th excavation parameter, w i is the weight factor, which indicates the importance of the i-th parameter to the objective function, and N is the total number of tunneling parameters that need to be optimized; Define the constraints and the calculation expression is as follows: P min ≤P i ≤P max , Where, P min and P max are the minimum and maximum values of the tunneling parameters, respectively.

8. The real-time dynamic optimization and adjustment method for the tunneling process according to claim 7, characterized in that: The Lagrange multiplier method is used to solve the problem. The calculation expression is as follows: ， Where, is the Lagrange function, λ is the Lagrange multiplier; Iteratively solve the optimal excavation parameters, and the calculation expression is as follows: ， Where, is the updated excavation parameter, α is the step size factor, It is the partial derivative of the objective function, which indicates the direction of influence of the current parameters on the optimization target; Adjust the tunnel boring machine parameters in real time to ensure the equipment operates in the optimal state and avoid resonance risks.

9. The real-time dynamic optimization and adjustment method for the tunneling process according to claim 1, characterized in that: The specific steps of dynamic feedback control are as follows: Adopting an adaptive neuro-fuzzy inference system for control optimization, the control rules can be adjusted according to the changes in the ground and the status of the equipment during the tunneling process; First, the historical excavation data is classified through the neural network self-learning mechanism, and the fuzzy control rules are adjusted according to the current excavation status; Then, fuzzy logic reasoning is used to evaluate the operating status of the tunnel boring machine in real time to determine whether the current tunneling parameters are in the resonance risk range; Finally, through fuzzy weight adjustment, the tunneling parameter adjustment strategy is continuously optimized to ensure the stability and efficiency of the tunneling system.

10. A real-time dynamic optimization and adjustment system for a tunneling process, for implementing the real-time dynamic optimization and adjustment method for a tunneling process according to any one of claims 1 to 9, characterized in that: It includes data acquisition and condition monitoring module, nonlinear dynamics modeling module, resonance risk prediction module, adaptive optimization and adjustment module, closed-loop feedback control module and abnormal state emergency processing module: The data acquisition and status monitoring module acquires the operating status data of the tunneling equipment in real time and records the dynamic change characteristics of the stratum; The nonlinear dynamic modeling module builds a nonlinear dynamic model of the tunneling process based on the collected data, uses an adaptive algorithm to identify the response characteristics of the formation, and establishes a mathematical relationship between the interaction between the equipment and the formation; The resonance risk prediction module uses a time series analysis method combined with the ground stiffness and vibration characteristics of the tunnel boring machine to calculate the resonance critical value under the current tunneling state, determine whether there is a low-frequency resonance trend, and warn of potential instability risks; The adaptive optimization and adjustment module dynamically adjusts key tunneling parameters based on prediction results and optimizes the tunneling mode through adaptive control algorithms, allowing equipment operation to avoid resonance zones while ensuring tunneling efficiency. A closed-loop feedback control module monitors in real time the impact of adjusted tunneling parameters on ground stability and equipment load. If the parameter adjustment still causes resonance, the tunneling strategy is further optimized to form a closed-loop control. The abnormal state emergency processing module automatically triggers the emergency mechanism when unpredictable formation mutations or abnormal equipment vibrations are detected, adjusts the excavation rhythm, reduces the cutterhead load, and optimizes the adjustment strategy based on historical data.

Images