Automatic jacking control system for tunnel jacking pipe

Abstract

The invention relates to the technical field of tunnel pipe jacking automatic jacking control, and discloses a tunnel pipe jacking automatic jacking control system which is characterized in that a sensor is used for detecting soil data in real time to obtain a soil data set, and b is the number of elements in the collected soil data set; according to the soil data set, a friction force prediction strategy is executed, and friction force during pipe jacking pushing is predicted; according to the actual thrust and the target thrust, an adaptive control strategy is executed, and the actual thrust is dynamically adjusted; the actual thrust obtained by executing the self-adaptive control strategy is applied to the virtual simulation model, the multi-dimensional evaluation strategy is executed, the optimal thrust is obtained, and the optimal thrust is applied to tunnel construction; according to the offset, a direction deviation correction strategy is executed, and the jacking direction and the actual thrust of the jacking pipe are corrected; and the fault alarm module is used for acquiring equipment parameters, executing a fault detection alarm strategy, detecting abnormal data and alarming the abnormal data. And more efficient and safer pipe jacking propulsion is realized.

Description

Technical Field

[0001] The invention relates to the technical field of tunnel pipe jacking automatic jacking control, in particular to a tunnel pipe jacking automatic jacking control system. Background Art

[0002] Pipe jacking technology is widely used in underground pipeline construction and tunnel construction, but in complex soil environments, there are many technical challenges in the process of advancement, mainly including soil friction, equipment failure, and directional deviation. These problems affect the efficiency of advancement and the stability of the equipment. In recent years, research has begun to introduce intelligent algorithms such as machine learning, deep learning, and reinforcement learning, combined with soil data for dynamic friction prediction and compensation. At the same time, modern intelligent deviation correction systems can automatically adjust the thrust direction according to the deviation during the advancement process with the help of real-time monitoring technologies such as inertial sensors, GPS, and lidar to ensure that the pipe jacking is advanced along the predetermined trajectory. In addition, health monitoring and fault diagnosis technologies have also begun to use big data analysis and intelligent algorithms, which can monitor the status of equipment in real time, automatically identify potential faults and issue early warnings, thereby improving the operational reliability of the equipment.

[0003] Traditional friction compensation methods and direction correction technologies often rely on static models or simple control algorithms, which cannot accurately respond to soil changes and dynamic errors during advancement, resulting in poor system adaptability and inability to adjust in real time to cope with different construction environments. In addition, equipment failures are often hidden, and traditional health monitoring systems rely on manual inspections or static thresholds, which cannot achieve real-time, intelligent fault diagnosis and repair, reducing equipment reliability and maintenance efficiency.

[0004] Therefore, the present invention proposes an automated tunnel jacking control system, which combines machine learning, virtual simulation and automated control technology to improve the intelligence, accuracy and reliability of the system, and ultimately achieve more efficient and safe tunnel jacking advancement. Summary of the invention

[0005] The present invention provides a tunnel pipe jacking automatic jacking control system, which helps solve the problems mentioned in the above background technology.

[0006] In a first aspect, the present application provides a tunnel jacking automatic jacking control system, which adopts the following technical solution: a tunnel jacking automatic jacking control system, comprising: Data acquisition module: Use sensors to detect soil data in real time and obtain the soil data set A(t) = {a1(t), a2(t), …, a b (t)}, b is the number of elements in the collected soil data set; Data processing module: Based on the soil data set, the friction prediction strategy is implemented to predict the friction force F1(t) when pushing the jacking pipe in the tunnel; According to the predicted friction force, the thrust compensation strategy is executed to obtain the target thrust F2(t) of the cylinder; Obtain the actual thrust F3(t) of pushing the jacking pipe in the tunnel; According to the actual thrust and target thrust, the adaptive control strategy is executed to dynamically adjust the actual thrust; Virtual simulation module: Build a virtual simulation model of tunnel construction; Apply the actual thrust obtained by executing the adaptive control strategy to the virtual simulation model, execute the multi-dimensional evaluation strategy, obtain the optimal thrust, and apply the optimal thrust to the tunnel construction; Correction module: During the tunnel construction process, real-time detection of the displacement of the jacking pipe; According to the offset, the direction correction strategy is implemented to correct the jacking direction and actual thrust of the jacking pipe; Fault alarm module: collects equipment parameters, executes fault detection and alarm strategies, detects abnormal data, and issues alarms for abnormal data.

[0007] Preferably, executing a friction force prediction strategy based on the soil data set to predict the friction force when pushing the jacking pipe in the tunnel includes: The soil data set is trained using a machine learning model to obtain the friction force F1(t) = f(a1(t), a2(t),…, a b (t)), where f(·) is the machine learning model; Obtain the actual friction force F1'(t) under the soil dataset and calculate the loss function of the machine learning model Use gradient descent to minimize the loss function and obtain the optimal regression coefficient; Set the deviation threshold ζ; If |F1'(t)-F1(t)|≤ζ, stop optimizing the regression coefficient.

[0008] By executing the friction prediction strategy and using the machine learning model to train the soil data set, the system can efficiently predict the friction of the tunnel jacking during the construction process. The advantage of this method is that it can accurately calculate the friction according to the actual soil conditions, thereby providing reliable data support for thrust compensation. The optimization of the machine learning model adjusts the regression coefficient through the gradient descent algorithm, so that the accuracy of friction prediction is continuously improved, and the error in the traditional prediction method is reduced. The setting of the deviation threshold further enhances the stability of the model, ensuring that the optimization is stopped in time when the prediction error is too large, avoiding the occurrence of overfitting. This prediction mechanism can be flexibly adjusted according to the actual situation, effectively ensuring the friction control during the construction process, and reducing the construction risks caused by improper thrust compensation.

[0009] Preferably, executing the thrust compensation strategy according to the predicted friction force to obtain the target thrust of the cylinder includes: Obtain the soil reaction force F0 when the jacking pipe pushes the soil; Calculate F2(t)=F0+F1(t).

[0010] Through the implementation of the thrust compensation strategy, combined with the friction prediction results, the target thrust of the cylinder can be calculated in real time, thereby accurately controlling the advancement process of the tunnel jacking pipe. This strategy can compensate for the impact caused by changes in soil friction by calculating and compensating the reaction force of the soil, so that the cylinder can stably output thrust under different soil conditions. The advantage of this strategy is that it avoids the situation where the thrust is too large or too small due to changes in soil friction during manual operation, thereby ensuring that the jacking pipe always remains on the predetermined advancement trajectory. Through precise control of the thrust, the equipment loss and time delay caused by improper thrust during the construction process can be effectively reduced, thereby improving the overall construction efficiency.

[0011] Preferably, executing an adaptive control strategy according to the actual thrust and the target thrust to dynamically adjust the actual thrust includes: Calculate the thrust error e(t), e(t) = F2(t) - F3(t); Set the PID gain parameters for dynamic optimization, namely proportional gain K1(t), which indicates the influence of the current error on the controller output, integral gain K2(t), which is used to eliminate the error, and differential gain K3(t), which is used to predict the future error trend; Calculate control signal Adaptive algorithm adjusts gain parameters; Where, K1(t)=K1(0)×(1+γ×e(t)); Among them, γ, λ is the adjustment coefficient, K1(0) is the initial proportional gain, K2(0) is the initial integral gain, and K3(0) is the initial differential gain; The actual thrust after adjustment is calculated as F4(t)=F3(t)+u(t).

[0012] The adaptive control strategy calculates the thrust error in real time and dynamically adjusts the PID gain parameters according to the error, which can make instant adjustments according to the actual changes in thrust during the construction process. The advantage of this strategy is that it can self-adjust according to the actual situation, avoiding the limitations of fixed gain parameters in traditional control methods, and ensuring the precise advancement of the jacking pipe under different construction environments. Adaptive control can not only effectively eliminate the thrust error, but also predict future error trends in real time, making the control system more adaptable and robust. By dynamically optimizing the PID parameters, the system can achieve precise thrust control, reduce the construction risks caused by unstable thrust, and maintain stable performance in long-term operation.

[0013] Preferably, the actual thrust obtained by executing the adaptive control strategy is applied to the virtual simulation model, and a multi-dimensional evaluation strategy is executed to obtain the optimal thrust, including: Calculate the thrust error e(t) = F2(t) - F3(t); Input the soil data set into the virtual simulation model to obtain the time t0 of the virtual simulation model response; Obtaining thrust fluctuations from a virtual simulation model The difference between actual thrust and friction: Calculate the multi-dimensional evaluation sum η1×e(t)+η2×t0+η3×o+η4×ΔF, where η1, η2, η3, η4 are thrust error weight, response time weight, thrust fluctuation weight and friction difference weight respectively; Set the number of candidates; The actual thrust corresponding to the sum of the multi-dimensional evaluations with the smallest number of candidates is obtained and recorded as the candidate thrust.

[0014] Establish a three-dimensional coordinate system in the tunnel; Get the position coordinates of each cylinder in, are the x-axis coordinate, y-axis coordinate and z-axis coordinate of the j-th cylinder in the three-dimensional coordinate system, j represents the j-th cylinder, 1≤j≤g1, g1 is the total number of cylinders; Let any candidate thrust be denoted as F total ; Calculate the energy efficiency of a single cylinder applying any thrust Among them, κ j is the efficiency factor of the cylinder, The spatial efficiency function for the thrust applied to the cylinder, is any component force of the candidate thrust, 1≤i≤g1; obtain the thrust of the cylinder Energy consumption Among them, ψ is a constant related to the cylinder efficiency; Calculate the objective function Candidate thrust meets The candidate thrust corresponding to the objective function is recorded as the optimal thrust F5(t) at time t.

[0015] By applying the actual thrust to the virtual simulation model and combining it with a multi-dimensional evaluation strategy, comprehensive thrust optimization can be performed before tunnel construction to ensure that the thrust during pipe jacking construction achieves the best effect. The advantage of this method is that it can not only simulate the thrust response under different soil conditions, but also further optimize the application of actual thrust by evaluating multiple factors such as thrust error, response time, and fluctuation. This multi-dimensional evaluation strategy provides a forward-looking solution that helps construction teams avoid construction delays or equipment damage caused by thrust mismatch in actual operations through simulation and calculation. At the same time, through continuous optimization and feedback adjustment, the system can effectively improve construction efficiency and thrust control accuracy and reduce construction risks.

[0016] Preferably, executing a direction correction strategy according to the offset to correct the jacking direction and actual thrust of the jacking pipe includes: The end of the jacking pipe that first contacts the soil is recorded as the first end; Record in real time the position of the first end in the three-dimensional coordinate system {x(t), y(t), z(t)}, where x(t) is the coordinate of the first end on the x-axis, y(t) is the coordinate of the first end on the y-axis, and z(t) is the coordinate of the first end on the z-axis; Get the target position coordinates {x0(t), y0(t), z0(t)} of the first end at time t; Calculate the offset between the current position and the target position to obtain the position error vector {Δx(t), Δy(t), Δz(t)}; The corrected actual thrust direction is the opposite direction of the position error vector; The change in actual thrust after correction is Where K4(t) is the correction gain coefficient; The actual thrust of the cylinder is calculated to be F6(t)=ΔF(t)+F5(t).

[0017] By detecting the offset of the jacking pipe in real time and executing the direction correction strategy, the thrust direction of the jacking pipe can be effectively corrected during the tunnel construction process to avoid deviation from the predetermined path. The advantage of this strategy is that it ensures the stable advancement of the jacking pipe under complex geological conditions by calculating the position error in real time and adjusting the direction, preventing excessive offset from affecting the construction progress and accuracy. By recording the movement trajectory of the jacking pipe in a three-dimensional coordinate system, the system can accurately determine the offset and adjust the thrust direction and size in time to ensure that the thrust direction during the construction process is always consistent with the target direction. This correction mechanism can greatly improve the accuracy and safety of jacking construction and avoid errors that may occur in the traditional manual correction process.

[0018] Preferably, executing a direction correction strategy according to the offset to correct the jacking direction and actual thrust of the jacking pipe includes: The self-correction strategy is implemented based on the reinforcement learning algorithm, specifically: Obtain the pushing speed v(t) of the jacking pipe at time t; Calculating the reward function in, The weight for balancing error and efficiency; Based on the reward function, the actual thrust is corrected again to F6(t) = -P({Δx(t),Δy(t),Δz(t)},v(t),{a1(t),a2(t),…,a b (t)}), where Ρ(·) is the strategy learned by the reinforcement learning algorithm; Get the maximum thrust F of the cylinder max ; The virtual simulation model is used to obtain all components of the total thrust F6(t), and each component is compared with the maximum thrust to obtain the component greater than the maximum thrust, which is recorded as the marked component. The distance that the oil cylinder applies the marked component force to push the jacking pipe is obtained and recorded as the marked distance; The position of the oil cylinder corresponding to the marked force component is obtained, and grouting is injected into the soil at the position so that the oil cylinder applies the maximum thrust to push the jacking pipe forward by a distance equal to the marked distance.

[0019] Through continuous learning and optimization strategies through the reinforcement learning algorithm, the actual thrust size can be automatically adjusted according to the trade-off between thrust error and efficiency. The advantage of this self-correcting strategy is that it can automatically learn the optimal thrust control strategy based on real-time construction data, avoiding errors and delays caused by human intervention. By calculating the reward function, the system can balance errors and efficiency to maximize the accuracy of thrust and construction efficiency. At the same time, the reinforcement learning algorithm can continuously adjust the strategy based on feedback to adapt to different soil conditions and construction environments, thereby improving the adaptability and intelligence of the system. This self-correcting strategy can significantly improve the level of automation in the construction process, reduce human intervention and operational errors, and improve the overall efficiency and safety of pipe jacking construction.

[0020] Preferably, the collecting device parameters, executing the fault detection and alarm strategy, detecting abnormal data, and issuing an alarm for the abnormal data include: Let any device parameter be q i , 1≤i≤g2, where g2 is the number of device parameters; For any device parameter, get the normal value q' of the device parameter i ; Detecting abnormal data, Δq i =q i -q i '; Get device parametersq i The fluctuation range Among them, q max is the device parameter q i The maximum value, q min is the device parameter q i The minimum value of Determine Δq i Is it exceeded If exceeded, the device parameter q is considered i is abnormal data; Provides alerts for abnormal data.

[0021] By collecting equipment parameters and executing fault detection and alarm strategies, the operating status of the equipment can be monitored in real time, abnormal data can be discovered in time, and alarms can be issued. The advantage of this strategy is that it can accurately determine whether there is equipment failure or abnormality by detecting the fluctuation range of equipment parameters, and issue alarms in time to avoid construction interruptions or safety accidents caused by equipment failure. The equipment fault detection mechanism can effectively improve the reliability and stability of equipment during the construction process, reducing the workload and errors of manual inspections. At the same time, real-time alarms enable operators to respond quickly and take necessary measures to repair or adjust, thereby ensuring the smooth progress of tunnel construction.

[0022] The present invention has the following beneficial effects: 1. The automatic jacking control system of the tunnel jacking uses sensors to detect soil data in real time, and executes friction prediction strategies based on these data. The system can accurately predict the friction during the tunnel jacking process. The benefits of this step are that, first of all, real-time acquisition of soil data can automatically adjust the construction strategy under different soil and environmental conditions to avoid uncertainty caused by changes in soil characteristics. The application of friction prediction strategies further optimizes the advancement effect of the jacking, making the friction prediction more accurate. This means that construction personnel can adjust the thrust compensation in advance during the actual construction process, thereby avoiding the failure of the jacking process due to excessive difficulty or insufficient thrust. In a data-driven way, the system can automatically adjust the operating parameters according to real-time feedback to improve the stability and efficiency of the construction process. This process not only avoids the errors that may be caused by manual intervention, but also effectively reduces equipment damage or construction delays caused by unreasonable friction estimation, greatly improving the reliability and safety of tunnel construction.

[0023] 2. The automatic jacking control system for tunnel jacking can automatically adjust the thrust compensation strategy according to the characteristics of the soil by using machine learning models to train soil data sets and optimize the friction prediction process. The main benefit is that through machine learning models, the system can continuously learn from historical data and improve its adaptability to different soil types. This makes friction prediction not only rely on traditional physical models, but can also dynamically optimize the prediction results in combination with big data analysis. This data-driven approach enables the model to accumulate more experience over time and maintain high prediction accuracy when facing complex or unknown soil conditions. The gradient descent optimization algorithm can effectively adjust the regression coefficient of the model, reduce prediction errors, and improve the overall performance of the system. In addition, after setting the deviation threshold, if the friction prediction has a large deviation, the system can stop the optimization in time to avoid overfitting and calculation errors. This mechanism can improve the stability and reliability of the model. The advantage of this step is that through accurate friction prediction, unnecessary energy waste and equipment loss are reduced, providing a more accurate basis for thrust compensation and optimizing the advancement efficiency of tunnel jacking.

[0024] 3. The automatic jacking control system of the tunnel jacking executes the thrust compensation strategy to obtain the target thrust of the cylinder according to the predicted friction. The benefit of this process is that it can ensure the coordination between the thrust and the soil friction. By dynamically adjusting the thrust of the cylinder, the system can effectively compensate for the changes caused by the soil friction and ensure that the jacking always maintains a stable thrust during the advancement process. Traditional tunnel construction often faces the problem of thrust mismatch, which may lead to equipment loss or uneven advancement progress. Through precise thrust compensation strategies, it can be ensured that the thrust changes are synchronized with the soil friction changes to avoid advancement obstacles caused by excessive or insufficient thrust. This can not only improve the stability and continuity of construction, but also reduce energy waste and equipment wear, and improve work efficiency. In addition, this strategy can also reduce equipment pressure fluctuations caused by rapid changes in friction, and improve the service life and stability of the jacking system.

[0025] 4. The tunnel jacking automatic control system dynamically adjusts the actual thrust through an adaptive control strategy, and the system can optimize the control parameters in real time according to the thrust error. The advantage of this step is that the system can automatically adjust the PID gain parameters to adapt to the thrust requirements in different construction environments, reducing the errors and delays of manual adjustments. In actual applications, the PID control algorithm can automatically adjust the proportional, integral and differential gains according to the current thrust error, thereby optimizing the thrust output in real time and ensuring that the system can cope with thrust fluctuations during construction. Adaptive control can not only improve the accuracy of the thrust, but also adjust the strategy according to actual construction feedback, making the thrust control more flexible and robust. In complex or uncertain construction environments, this adaptive capability enables the system to operate stably, avoiding the instability caused by traditional fixed parameter control methods. In addition, real-time optimization of gain parameters can also reduce energy waste and thrust errors, ensure smooth construction, and improve construction efficiency.

[0026] 5. The automatic jacking control system of the tunnel jacking system can simulate different thrust schemes and select the optimal thrust before actual construction by building a virtual simulation model and applying a multi-dimensional evaluation strategy. The benefit of this step is that the virtual simulation can provide thrust response predictions under different soil conditions without interfering with the actual construction. By calculating the thrust error, response time, thrust fluctuation and friction difference, the system can comprehensively evaluate the effects of different thrust strategies and select the optimal thrust that minimizes the comprehensive evaluation. This evaluation strategy can effectively avoid the risks caused by improper thrust selection in actual construction and ensure the maximization of construction progress and accuracy. Through virtual simulation, the construction team can foresee and solve possible problems before construction, thereby optimizing the thrust application in advance and reducing the need for on-site adjustments. The introduction of a multi-dimensional evaluation strategy allows thrust adjustment to not only take into account a single factor, but also maximize the stability and efficiency of the construction process through comprehensive analysis from multiple angles.

[0027] 6. The automatic jacking control system for tunnel jacking can detect the offset of the jacking pipe in real time and correct the thrust direction through the direction correction strategy, so as to ensure that the jacking pipe always advances along the predetermined trajectory. The advantage of this step is that the system can monitor the position of the jacking pipe in real time, promptly detect and correct the offset problem, thereby preventing the jacking pipe from deviating from the designed track during the construction process. This offset detection and correction strategy accurately calculates the position of the jacking pipe through a three-dimensional coordinate system, and makes real-time adjustments according to the target position to ensure accurate correction of the thrust direction. By reversely correcting the position error vector, the system can dynamically adjust the thrust direction to avoid construction problems caused by excessive offset. Real-time correction not only improves the accuracy of construction, but also reduces equipment damage or construction delays caused by error accumulation, ensuring the smooth progress of the project. In addition, the automation of the correction process reduces the need for manual intervention and improves the safety and accuracy of operations.

[0028] 7. The tunnel jacking automatic control system, based on the self-correction strategy of the reinforcement learning algorithm, can more effectively balance the error and construction efficiency by continuously optimizing the thrust adjustment. The benefit of this step is that reinforcement learning can automatically learn and adjust the correction strategy according to real-time construction data, so that the system can adapt to different construction environments and soil conditions. The reinforcement learning model can guide the system to maintain construction efficiency while minimizing the error through the reward function, thereby optimizing the accuracy of thrust adjustment. Compared with the traditional empirical adjustment method, reinforcement learning can continuously improve the decision-making strategy through multiple iterations and enhance the adaptive ability of the system. Through the intelligent learning process, the system can adjust the thrust in real time, maintain the propulsion stability of the jacking pipe, reduce manual intervention in complex environments, and improve construction efficiency and accuracy. In addition, this self-correction strategy provides the system with stronger intelligence and automation, further promoting the intelligent process of tunnel construction.

[0029] 8. The automatic jacking control system of the tunnel jacking pipe obtains all the components of the total thrust during the jacking pipe construction process through the virtual simulation model, and compares each component with the maximum thrust of the cylinder. It can accurately identify the components that exceed the thrust range of the cylinder, and these components are recorded as marked components. By obtaining the position information of the cylinder when applying the marked component, the area where the soil bearing pressure is too large can be accurately located. By grouting the soil at these marked positions, the mechanical properties of the soil can be effectively improved, the local friction and soil resistance can be reduced, and the propulsion conditions of the pipeline can be optimized. In this way, the cylinder can propel the jacking pipe with the maximum thrust, so that the propulsion distance reaches the expected marked distance, and the construction efficiency is improved. By controlling the matching relationship between the thrust and the soil conditions, the mechanical load of the equipment can be reduced, the equipment life can be extended, and the risk of jacking pipe displacement or jamming during the propulsion process can be reduced. By optimizing the construction parameters and soil conditions, it can not only save energy and construction costs, but also effectively reduce the impact of construction on the surrounding environment, and improve the safety and economy of the overall construction. BRIEF DESCRIPTION OF THE DRAWINGS

[0030] Figure 1 It is a schematic diagram of the process of the present invention.

[0031] Figure 2 It is a schematic diagram of the system module of the present invention. DETAILED DESCRIPTION

[0032] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0033] Embodiment 1, refer to Figure 1 , a tunnel jacking automatic jacking control system, including: a data acquisition module: using sensors to detect soil data in real time, obtaining a soil data set A(t) = {a1(t), a2(t), …, a b (t)}, b is the number of elements in the collected soil data set; In this embodiment, the soil data collected in real time includes humidity, temperature, pressure, etc.; Data processing module: Based on the soil data set, the friction prediction strategy is implemented to predict the friction force F1(t) when pushing the jacking pipe in the tunnel; According to the predicted friction force, the thrust compensation strategy is executed to obtain the target thrust F2(t) of the cylinder; Obtain the actual thrust F3(t) of pushing the jacking pipe in the tunnel; According to the actual thrust and target thrust, the adaptive control strategy is executed to dynamically adjust the actual thrust; Virtual simulation module: Build a virtual simulation model of tunnel construction; Apply the actual thrust obtained by executing the adaptive control strategy to the virtual simulation model, execute the multi-dimensional evaluation strategy, obtain the optimal thrust, and apply the optimal thrust to the tunnel construction; Correction module: During the tunnel construction process, real-time detection of the displacement of the jacking pipe; According to the offset, the direction correction strategy is implemented to correct the jacking direction and actual thrust of the jacking pipe; Fault alarm module: collects equipment parameters, executes fault detection and alarm strategies, detects abnormal data, and issues alarms for abnormal data.

[0034] In this embodiment, the module diagram refers to Figure 2 .

[0035] According to the soil data set, a friction prediction strategy is executed to predict the friction when pushing the jacking pipe in the tunnel, including: using a machine learning model to train the soil data set to obtain the friction F1(t) = f(a1(t), a2(t), …, a b (t)), where f(·) is the machine learning model; Obtain the actual friction force F1'(t) under the soil dataset and calculate the loss function of the machine learning model Use gradient descent to minimize the loss function and obtain the optimal regression coefficient; Set the deviation threshold ζ; If |F1'(t)-F1(t)|≤ζ, stop optimizing the regression coefficient.

[0036] By using sensors to monitor soil data in real time and perform friction prediction, the system can dynamically adjust the thrust according to the real-time soil conditions. The benefit of this process is that through accurate friction prediction, the system can respond to changes in soil friction in advance, avoiding construction instability caused by excessive or insufficient friction during pipe jacking. Through data collection and analysis, the system can identify the different characteristics of the soil and accurately calculate friction, thereby providing a reliable basis for thrust compensation. Compared with traditional methods, real-time feedback prediction can reduce uncertainty in construction and improve construction accuracy and efficiency. In addition, the system can quickly adjust the thrust compensation strategy according to changes in soil conditions, effectively reducing equipment wear and energy waste, and extending equipment life.

[0037] By training and optimizing the soil data set through machine learning models, the system can adjust the accuracy of friction prediction in real time. The benefit of this method is that through continuous optimization of machine learning, the system can adapt to complex soil environments and provide more accurate friction predictions. Through the gradient descent optimization algorithm, the model can reduce prediction errors, thereby ensuring more accurate thrust compensation. In addition, by setting the deviation threshold, the system can avoid overfitting and maintain the stability of the model. Compared with traditional static models, machine learning methods are more adaptable and flexible, and can provide reliable friction predictions in different construction environments, reducing construction problems caused by soil changes. Ultimately, this will help to achieve precise control of pipe jacking construction and reduce construction risks and thrust mismatches.

[0038] According to the predicted friction force, the thrust compensation strategy is executed to obtain the target thrust of the cylinder, including: Obtain the soil reaction force F0 when the jacking pipe pushes the soil; Calculate F2(t)=F0+F1(t).

[0039] Through the thrust compensation strategy, the system can adjust the target thrust of the cylinder according to the friction prediction results to ensure the stable advancement of the jacking pipe. The advantage of this strategy is that the system can calculate and adjust the thrust in real time to adapt to the changes in friction of different soils, avoiding the problem of excessive or insufficient thrust. Through dynamic compensation, the system ensures the coordination between thrust and friction, thereby improving the smoothness and efficiency of advancement. Compared with traditional manual adjustment methods, automated thrust compensation can reduce the risk of human errors and untimely adjustments, ensure that the jacking pipe is always advanced along the predetermined path, and avoid equipment damage or progress delays caused by thrust mismatch during construction. Ultimately, this not only improves construction accuracy, but also greatly improves construction efficiency and safety.

[0040] According to the actual thrust and target thrust, the adaptive control strategy is executed to dynamically adjust the actual thrust, including: Calculate the thrust error e(t), e(t) = F2(t) - F3(t); Set the PID gain parameters for dynamic optimization, namely proportional gain K1(t), which indicates the influence of the current error on the controller output, integral gain K2(t), which is used to eliminate the error, and differential gain K3(t), which is used to predict the future error trend; Calculate control signal Adaptive algorithm adjusts gain parameters; Where, K1(t)=K1(0)×(1+γ×e(t)); Among them, γ, λ is the adjustment coefficient, K1(0) is the initial proportional gain, K2(0) is the initial integral gain, and K3(0) is the initial differential gain; The actual thrust after adjustment is calculated as F4(t)=F3(t)+u(t).

[0041] Through the adaptive control strategy, the system can dynamically adjust the control parameters according to the error between the actual thrust and the target thrust. The benefit of this process is that the system can optimize the PID gain parameters in real time to ensure that the thrust control adapts to different construction environments and soil conditions. Through the dynamic adjustment of proportional, integral and differential gains, the system can reduce the thrust error and improve the control accuracy. Compared with the traditional fixed parameter control method, adaptive control can automatically adjust according to real-time feedback, avoiding the lag and error of manual adjustment, so that the jacking pipe can operate stably under various construction conditions. In addition, this method can also improve the robustness of the system, reduce the impact of external disturbances on the construction process, and ultimately improve the construction efficiency and stability of the construction.

[0042] Apply the actual thrust obtained by executing the adaptive control strategy to the virtual simulation model and execute a multi-dimensional evaluation strategy to obtain the optimal thrust, including: Calculate the thrust error e(t) = F2(t) - F3(t); Input the soil data set into the virtual simulation model to obtain the time t0 of the virtual simulation model response; Obtaining thrust fluctuations from a virtual simulation model The difference between actual thrust and friction: Calculate the multi-dimensional evaluation sum η1×e(t)+η2×t0+η3×o+η4×ΔF, where η1, η2, η3, η4 are thrust error weight, response time weight, thrust fluctuation weight and friction difference weight respectively; Set the number of candidates; Obtain the actual thrust corresponding to the sum of the multi-dimensional evaluations with the smallest number of candidates, recorded as the candidate thrust. Establish a three-dimensional coordinate system in the tunnel; Get the position coordinates of each cylinder in, are the x-axis coordinate, y-axis coordinate and z-axis coordinate of the j-th cylinder in the three-dimensional coordinate system, j represents the j-th cylinder, 1≤j≤g1, g1 is the total number of cylinders; Let any candidate thrust be denoted as F total ; Calculate the energy efficiency of a single cylinder applying any thrust Among them, κ j is the efficiency factor of the cylinder, The spatial efficiency function for the thrust applied to the cylinder, is any component force of the candidate thrust, 1≤i≤g1; obtain the thrust of the cylinder Energy consumption Among them, ψ is a constant related to the cylinder efficiency; Calculate the objective function Candidate thrust meets The candidate thrust corresponding to the objective function is recorded as the optimal thrust F5(t) at time t.

[0043] By applying the actual thrust to the virtual simulation model and executing a multi-dimensional evaluation strategy, the system can simulate the thrust effect under different soil conditions and select the optimal thrust. The benefit of this strategy is that virtual simulation can predict the effect of thrust adjustment in advance, avoiding the thrust mismatch problem that may occur during the actual construction process. By evaluating multiple factors such as thrust error, response time, thrust fluctuation, etc., the system can comprehensively optimize the thrust so that it can achieve the best effect under different conditions. This multi-dimensional evaluation strategy enables the system to comprehensively consider various construction factors, thereby avoiding the adverse effects caused by a single factor and maximizing construction efficiency and accuracy. Through virtual simulation, the construction team can be fully prepared before actual operation, reducing the need for on-site adjustments and ensuring the smooth progress of the construction process.

[0044] According to the offset, the direction correction strategy is implemented to correct the jacking direction and actual thrust of the jacking pipe, including: The end of the jacking pipe that first contacts the soil is recorded as the first end; Record in real time the position of the first end in the three-dimensional coordinate system {x(t), y(t), z(t)}, where x(t) is the coordinate of the first end on the x-axis, y(t) is the coordinate of the first end on the y-axis, and z(t) is the coordinate of the first end on the z-axis; Get the target position coordinates {x0(t), y0(t), z0(t)} of the first end at time t; Calculate the offset between the current position and the target position to obtain the position error vector {Δx(t), Δy(t), Δz(t)}; The corrected actual thrust direction is the opposite direction of the position error vector; The change in actual thrust after correction is Where K4(t) is the correction gain coefficient; The actual thrust of the cylinder is calculated to be F6(t)=ΔF(t)+F5(t).

[0045] By detecting the offset of the jacking pipe in real time and executing the direction correction strategy, the system can ensure that the jacking pipe always advances along the predetermined trajectory. The advantage of this strategy is that real-time correction can prevent the jacking pipe from deviating from the target path due to errors, ensuring the accuracy and safety of construction. By accurately calculating the position error through the three-dimensional coordinate system, the system can dynamically adjust the thrust direction and size to correct the offset. Compared with manual intervention, automated correction has higher accuracy and response speed, and can correct the offset in real time during the construction process, avoiding construction problems caused by error accumulation. Through this automated correction mechanism, the construction team can complete tasks more efficiently, reduce equipment damage and project delays caused by offsets, and ultimately improve construction efficiency and safety.

[0046] According to the offset, the direction correction strategy is implemented to correct the jacking direction and actual thrust of the jacking pipe, including: The self-correction strategy is implemented based on the reinforcement learning algorithm, specifically: Obtain the pushing speed v(t) of the jacking pipe at time t; Calculating the reward function in, The weight for balancing error and efficiency; Based on the reward function, the actual thrust is corrected again to F6(t) = -P({Δx(t),Δy(t),Δz(t)},v(t),{a1(t),a2(t),…,a b (t)}), where Ρ(·) is the strategy learned by the reinforcement learning algorithm; Get the maximum thrust F of the cylinder max ; The virtual simulation model is used to obtain all components of the total thrust F6(t), and each component is compared with the maximum thrust to obtain the component greater than the maximum thrust, which is recorded as the marked component. The distance that the oil cylinder applies the marked component force to push the jacking pipe is obtained and recorded as the marked distance; The position of the oil cylinder corresponding to the marked force component is obtained, and grouting is injected into the soil at the position so that the oil cylinder applies the maximum thrust to push the jacking pipe forward by a distance equal to the marked distance.

[0047] By obtaining all the components of the total thrust during the pipe jacking construction process through the virtual simulation model and comparing each component with the maximum thrust of the cylinder, the components that exceed the thrust range of the cylinder can be accurately identified, and these components are recorded as marked components. By obtaining the position information of the cylinder when applying the marked component, the area where the soil bearing pressure is too large can be accurately located. By grouting the soil at these marked positions, the mechanical properties of the soil can be effectively improved, the local friction and soil resistance can be reduced, and the propulsion conditions of the pipeline can be optimized. In this way, the cylinder can propel the pipe jacking with the maximum thrust, so that the propulsion distance reaches the expected marked distance, and the construction efficiency is improved. By controlling the matching relationship between the thrust and the soil conditions, the mechanical load of the equipment can be reduced, the equipment life can be extended, and the risk of pipe jacking displacement or jamming during the propulsion process can be reduced. By optimizing the construction parameters and soil conditions, not only can energy and construction costs be saved, but also the impact of construction on the surrounding environment can be effectively reduced, and the safety and economy of the overall construction can be improved.

[0048] Through the reinforcement learning algorithm, the system can self-learn and adjust the thrust compensation strategy to achieve optimal self-correction. The advantage of this strategy is that reinforcement learning can automatically optimize the thrust adjustment plan based on actual construction data, thereby reducing human intervention. By balancing errors and efficiency through the reward function, the system can find the best thrust correction strategy to adapt to different construction environments. Compared with traditional fixed strategies, reinforcement learning can continuously optimize the thrust control plan through continuous iterative adjustments. The system can not only automatically learn and adjust during the construction process, but also cope with changes in complex environments and maintain the accuracy and stability of construction. Ultimately, the self-correction strategy based on reinforcement learning can improve the intelligence level of the system, reduce human errors, and improve the degree of automation and efficiency of construction.

[0049] Collect equipment parameters, execute fault detection and alarm strategies, detect abnormal data, and issue alarms for abnormal data, including: Let any device parameter be q i , 1≤i≤g2, where g2 is the number of device parameters; In this embodiment, the equipment parameters include cylinder temperature, cylinder working pressure, cylinder displacement, cylinder displacement speed (thrust change rate), cylinder acceleration, motor speed and equipment vibration acceleration.

[0050] For any device parameter, get the normal value q' of the device parameter i ; Detecting abnormal data, Δq i =q i -q i '; Get device parametersq i The fluctuation range Among them, qmax is the device parameter q i The maximum value, q min is the device parameter q i The minimum value of Determine Δq i Is it exceeded If exceeded, the device parameter q is considered i is abnormal data; Provides alerts for abnormal data.

[0051] By collecting equipment parameters and executing fault detection and alarm strategies, the system can monitor the operating status of the equipment in real time and detect potential faults in a timely manner. The advantage of this strategy is that the system can quickly determine whether there is any abnormality in the equipment by detecting the fluctuation range of equipment parameters and automatically issue an alarm. By discovering problems in advance, the construction team can take timely measures to avoid construction delays or safety accidents caused by equipment failure. Compared with traditional manual detection, automated fault detection can respond in real time, reducing the risk of manual errors and inspection lags. This strategy can also effectively reduce equipment maintenance costs, increase the service life of equipment, and ensure that the equipment is always in the best operating state during the construction process, thereby improving the stability and safety of construction.

[0052] It should be noted that, in this article, relational terms such as first and second, etc. 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 "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including 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.

[0053] The above are only preferred embodiments of the present invention. It should be pointed out that, for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the technical principles of the present invention. These improvements and modifications should also be regarded as within the scope of protection of the present invention.

Claims

1. A tunnel jacking automatic jacking control system, characterized in that: include: Data acquisition module: Use sensors to detect soil data in real time and obtain the soil data set A(t) = {a1(t), a2(t), …, a b (t)}, b is the number of elements in the collected soil data set; Data processing module: Based on the soil data set, the friction prediction strategy is implemented to predict the friction force F1(t) when pushing the jacking pipe in the tunnel; According to the predicted friction force, the thrust compensation strategy is executed to obtain the target thrust F2(t) of the cylinder; Obtain the actual thrust F3(t) of pushing the jacking pipe in the tunnel; According to the actual thrust and target thrust, the adaptive control strategy is executed to dynamically adjust the actual thrust; Virtual simulation module: Construct a virtual simulation model of tunnel construction; Apply the actual thrust obtained by executing the adaptive control strategy to the virtual simulation model, execute the multi-dimensional evaluation strategy, obtain the optimal thrust, and apply the optimal thrust to the tunnel construction; Correction module: During the tunnel construction process, real-time detection of the displacement of the jacking pipe; According to the offset, the direction correction strategy is implemented to correct the jacking direction and actual thrust of the jacking pipe; Fault alarm module: collects equipment parameters, executes fault detection and alarm strategies, detects abnormal data, and issues alarms for abnormal data.

2. The tunnel jacking automatic jacking control system according to claim 1 is characterized in that: The method of executing a friction force prediction strategy based on the soil data set to predict the friction force when the jacking pipe is pushed in the tunnel includes: The soil data set is trained using a machine learning model to obtain the friction force F1(t) = f(a1(t), a2(t),…, a b (t)), where f(·) is the machine learning model; Obtain the actual friction force F1'(t) under the soil dataset and calculate the loss function of the machine learning model Use gradient descent to minimize the loss function and obtain the optimal regression coefficient; Set the deviation threshold ζ; If |F1'(t)-F1(t)|≤ζ, stop optimizing the regression coefficient.

3. The tunnel jacking automatic jacking control system according to claim 2 is characterized in that: The method of executing the thrust compensation strategy according to the predicted friction force to obtain the target thrust of the cylinder includes: Obtain the soil reaction force F0 when the jacking pipe pushes the soil; Calculate F2(t)=F0+F1(t).

4. The tunnel jacking automatic jacking control system according to claim 1 is characterized in that: The method of executing an adaptive control strategy according to the actual thrust and the target thrust to dynamically adjust the actual thrust includes: Calculate the thrust error e(t), e(t) = F2(t) - F3(t); Set the PID gain parameters for dynamic optimization, namely proportional gain K1(t), which indicates the influence of the current error on the controller output, integral gain K2(t), which is used to eliminate the error, and differential gain K3(t), which is used to predict the future error trend; Calculate control signal Adaptive algorithm adjusts gain parameters; Where, K1(t)=K1(0)×(1+γ×e(t)); Among them, γ, λ is the adjustment coefficient, K1(0) is the initial proportional gain, K2(0) is the initial integral gain, and K3(0) is the initial differential gain; The actual thrust after adjustment is calculated as F4(t)=F3(t)+u(t).

5. The tunnel jacking automatic jacking control system according to claim 4 is characterized in that: The actual thrust obtained by executing the adaptive control strategy is applied to the virtual simulation model, and a multi-dimensional evaluation strategy is executed to obtain the optimal thrust, including: Calculate the thrust error e(t) = F2(t) - F3(t); Input the soil data set into the virtual simulation model to obtain the time t0 of the virtual simulation model response; Obtaining thrust fluctuations from a virtual simulation model The difference between actual thrust and friction: Calculate the multi-dimensional evaluation sum η1×e(t)+η2×t0+η3×o+η4×ΔF, where η1, η2, η3, η4 are thrust error weight, response time weight, thrust fluctuation weight and friction difference weight respectively; Set the number of candidates; The actual thrust corresponding to the sum of the multi-dimensional evaluations with the smallest number of candidates is obtained and recorded as the candidate thrust.

6. The tunnel jacking automatic jacking control system according to claim 5 is characterized in that: The actual thrust obtained by executing the adaptive control strategy is applied to the virtual simulation model, and the multi-dimensional evaluation strategy is executed to obtain the optimal thrust, and further includes: establishing a three-dimensional coordinate system in the tunnel; Get the position coordinates of each cylinder in, are the x-axis coordinate, y-axis coordinate and z-axis coordinate of the j-th cylinder in the three-dimensional coordinate system, j represents the j-th cylinder, 1≤j≤g1, g1 is the total number of cylinders; Let any candidate thrust be F total ; Calculate the energy efficiency of a single cylinder applying any thrust Among them, κ j is the efficiency factor of the cylinder, The spatial efficiency function for the thrust applied to the cylinder, is any component force of the candidate thrust, 1≤i≤g1; obtain the thrust of the cylinder Energy consumption Among them, ψ is a constant related to the cylinder efficiency; Calculate the objective function Candidate thrust meets The candidate thrust corresponding to the objective function is recorded as the optimal thrust F5(t) at time t.

7. The tunnel jacking automatic jacking control system according to claim 6 is characterized in that: The direction correction strategy is executed according to the offset to correct the jacking direction and actual thrust of the jacking pipe, including: The end of the jacking pipe that first contacts the soil is recorded as the first end; Record in real time the position of the first end in the three-dimensional coordinate system {x(t), y(t), z(t)}, where x(t) is the coordinate of the first end on the x-axis, y(t) is the coordinate of the first end on the y-axis, and z(t) is the coordinate of the first end on the z-axis; Get the target position coordinates {x0(t), y0(t), z0(t)} of the first end at time t; Calculate the offset between the current position and the target position to obtain the position error vector {Δx(t), Δy(t), Δz(t)}; The corrected actual thrust direction is the opposite direction of the position error vector; The change in actual thrust after correction is Where K4(t) is the correction gain coefficient; The actual thrust of the cylinder is calculated to be F6(t)=ΔF(t)+F5(t).

8. The tunnel jacking automatic jacking control system according to claim 7 is characterized in that: The direction correction strategy is executed according to the offset to correct the jacking direction and actual thrust of the jacking pipe, including: The self-correction strategy is implemented based on the reinforcement learning algorithm, specifically: Obtain the pushing speed v(t) of the jacking pipe at time t; Calculating the reward function in, The weight for balancing error and efficiency; Based on the reward function, the actual thrust is corrected again to F6(t) = -P({Δx(t),Δy(t),Δz(t)},v(t),{a1(t),a2(t),…,a b (t)}), where Ρ(·) is the strategy learned by the reinforcement learning algorithm; Get the maximum thrust F of the cylinder max ; The virtual simulation model is used to obtain all components of the total thrust F6(t), and each component is compared with the maximum thrust to obtain the component greater than the maximum thrust, which is recorded as the marked component. The distance that the oil cylinder applies the marked component force to push the jacking pipe is obtained and recorded as the marked distance; The position of the oil cylinder corresponding to the marked force component is obtained, and grouting is injected into the soil at the position so that the oil cylinder applies the maximum thrust to push the jacking pipe forward by a distance equal to the marked distance.

9. The tunnel jacking automatic jacking control system according to claim 1, characterized in that: The collecting of equipment parameters, executing fault detection and alarm strategies, detecting abnormal data, and issuing alarms for abnormal data include: Let any device parameter be q i , 1≤i≤g2, where g2 is the number of device parameters; For any device parameter, get the normal value q' of the device parameter i ; Detecting abnormal data, Δq i =q i -q i '; Get device parametersq i The fluctuation range Among them, q max is the device parameter q i The maximum value, q min For Reserve parameter q i The minimum value of Determine Δq i Is it exceeded If exceeded, the device parameter q is considered i is abnormal data; Issue alerts on abnormal data.

Images