Shield synchronous pushing and splicing oil cylinder thrust determining method and system, medium and equipment

Abstract

The invention relates to the technical field of shield tunnel construction, and provides a shield synchronous pushing and splicing oil cylinder thrust determining method and system, a medium and equipment, and the method comprises the steps that after a shield tunneling machine enters a synchronous pushing and splicing state, a ring of duct pieces are divided into oil cylinder partitions according to the number and positions of the duct pieces, each duct piece corresponds to one partition, and the oil cylinder pressure in each partition is the same; obtaining observation data at a historical moment and pressure of each oil cylinder at a future moment, and predicting attitude deviation at the future moment through the tunneling attitude deviation prediction model; the total propulsive force, the resultant moment in the horizontal direction and the resultant moment in the vertical direction are obtained, the attitude deviation at the future moment is combined, and the oil cylinder thrust of each partition when the minimum energy consumption cost and the minimum attitude deviation cost are met is solved through a model prediction control framework; through a pre-synchronization method, the oil cylinder thrust of each subarea is smoothly adjusted before state switching. The method has strong adaptability and fault tolerance, and is suitable for engineering application in a complex geological environment.

Description

Technical Field

[0001] The invention belongs to the technical field of shield tunnel construction, and particularly relates to a method, a system, a medium and a device for determining the thrust of a shield synchronous pushing and assembling oil cylinder. Background Technique

[0002] The statements in this part only provide background technical information related to the present invention, and do not necessarily constitute prior art.

[0003] With the advantages of fast construction speed, strong safety, high automation degree and little impact on the surrounding environment, the shield method has been widely used in the field of tunnel construction. With the increase of traffic demand and the development of construction technology, more and more long-distance tunnel projects have emerged, putting forward higher requirements for the construction efficiency of shield tunneling. The conventional shield tunneling construction adopts a series of construction procedures of "advancing - stopping - assembling". The time required for the shield machine to advance forward and the segment assembly is similar. The current construction status of separating pushing and assembling not only seriously restricts the construction efficiency, but also requires frequent machine replacement during the construction process, which is likely to increase the risk of equipment failure, aggravate the ground settlement, and even cause ground collapse, resulting in adverse economic and social impacts.

[0004] To shorten the construction progress and improve the construction efficiency, the synchronous pushing and assembling technology has emerged, and the tunneling and assembling tasks are carried out simultaneously. Specifically, during the shield tunneling construction process, according to the current segment assembly position, the corresponding pushing cylinders are retracted. To ensure that the shield tunnels normally along the designed axis, the thrust of the remaining cylinders is redistributed to ensure that the resultant force and the resultant moment are the same as those in the single tunneling state.

[0005] The key to the shield synchronous pushing and assembling technology lies in regulating the thrust of the hydraulic cylinders to ensure force balance and enabling the shield machine to tunnel along the designed axis of the tunnel. Especially when some cylinders in the propulsion system are absent, how to reasonably distribute the total thrust to the remaining cylinders to ensure that the resultant force can balance the external load of the shield machine and tunnel along the designed axis at the same time.

[0006] At present, the research on thrust distribution focuses on distributing the resultant thrust of the absent cylinders to the remaining cylinders by means of incremental distribution or gradient distribution. This method can only ensure that the magnitude and action point of the total thrust remain unchanged, without considering the effect of the shield pose. Unreasonable cylinder thrust will damage the propulsion mechanism, resulting in phenomena such as segment eccentric loading. With the in-depth research, some scholars have realized the decision-making of cylinder component forces by establishing a constraint optimization model, but less consideration has been given to the possible impact on the shield attitude due to the retraction of some cylinders.

[0007] Model Predictive Control (MPC) is an advanced control strategy based on dynamic models, rolling optimization, and feedback correction. It is mainly used for the real-time control of complex dynamic systems and can well solve the problem of cylinder thrust decision-making during the synchronous pushing and assembling process. It can not only reasonably distribute the cylinder thrust but also minimize the attitude deviation as much as possible to achieve the closed-loop control of the synchronous pushing and assembling process. Summary of the Invention

[0008] To solve the technical problems existing in the above-mentioned background technology, the present invention provides a method, system, medium, and device for determining the thrust of shield synchronous pushing cylinders. It adopts a segment-cylinder dynamic zoning mechanism to divide the cylinder zones in real time according to the actual assembling requirements, ensuring the pressure balance of the cylinders in each zone under different assembling forms and having strong dynamic adaptability. It obtains the total propulsion force, the resultant moment in the horizontal direction, the resultant moment in the vertical direction, and other key construction and geological parameters, establishes a tunneling attitude deviation prediction model and embeds it into the model predictive control framework, and solves the cylinder thrust distribution strategy for different zones on the premise of meeting the minimum energy consumption cost and attitude deviation cost to ensure the coordinated control of the shield synchronous pushing.

[0009] To achieve the above object, the present invention adopts the following technical solutions: The first aspect of the present invention provides a method for determining the thrust of shield synchronous pushing cylinders, which includes: After the shield machine enters the synchronous pushing state, divide the cylinder zones according to the number and position of segments in one ring. Each segment corresponds to one zone, and the pressure of the cylinders within each zone is the same; Obtain the observed data at the historical moment and the pressure of each cylinder at the future moment, and predict the attitude deviation at the future moment through the tunneling attitude deviation prediction model; obtain the total propulsion force, the resultant moment in the horizontal direction, and the resultant moment in the vertical direction, and combine the attitude deviation at the future moment to solve the cylinder thrust of each zone through the model predictive control framework when meeting the minimum energy consumption cost and attitude deviation cost; Through the pre-synchronization method, smoothly adjust the cylinder thrust of each zone before the state switch.

[0010] Further, the tunneling attitude deviation prediction model includes two branches: Branch one takes the observed data at the historical moment as the input, adopts a CNN-Transformer joint encoder to obtain the global state representation vector; Branch two takes the pressure of each cylinder at the future moment as the input, adopts a two-layer fully connected network encoding to obtain the operation intention vector; after the global state representation vector and the operation intention vector are fused through feature splicing and residual mapping, they are non-linearly mapped through a two-layer fully connected network to obtain the attitude deviation at the future moment.

[0011] Furthermore, for the tunneling attitude deviation prediction model, the observed data at the historical moment includes the shield head horizontal deviation, shield head vertical deviation, shield tail horizontal deviation, shield tail vertical deviation, actual thrust of each propulsion cylinder, propulsion speed, cutter head torque, cutter head rotation speed, and formation type; The attitude deviations at the future moment include the shield head horizontal deviation, shield head vertical deviation, shield tail horizontal deviation, and shield tail vertical deviation.

[0012] Furthermore, during the solving process of the cylinder thrust of each partition, constraints are imposed on the cylinder thrust of each partition and the attitude deviation.

[0013] Furthermore, the prediction model adopted during the solving process of the cylinder thrust of each partition is expressed as:

[0014]

[0015] where r represents the radius of the hydraulic cylinder distribution circle, is the angle between the first cylinder in the first partition (zone A) and the x-axis, is the angle between any two adjacent cylinders, represents the force magnitude of each propulsion cylinder in the i-th partition, represents the propulsion cylinders included in the i-th partition, and k represents the cylinder number.

[0016] Furthermore, during the solving process of the cylinder thrust of each partition, a rolling time domain strategy is adopted, and the particle swarm optimization algorithm is used to traverse all feasible solution spaces.

[0017] Furthermore, during the solving process of the cylinder thrust of each partition, the actual state of the shield machine is monitored in real time. The actual measured values are compared with the output of the prediction model, the error is calculated, and based on the error, the weights of the cost function of the model predictive control framework are corrected.

[0018] The second aspect of the present invention provides a system for determining the thrust of synchronous pushing and assembling cylinders of a shield machine, which includes: A partition module, which is configured to: after the shield machine enters the synchronous pushing and assembling state, divide a ring of segments into cylinder partitions according to the number and position of the segments, each segment corresponds to one partition, and the cylinder pressures within each partition are the same; A cylinder thrust determination module, which is configured to: obtain the observed data at the historical moment and the cylinder pressures at the future moment, predict the attitude deviations at the future moment through the tunneling attitude deviation prediction model; obtain the total propulsion force, the resultant moment in the horizontal direction, and the resultant moment in the vertical direction, and combine the attitude deviations at the future moment, and through the model predictive control framework, solve the cylinder thrust of each partition when the minimum energy consumption cost and attitude deviation cost are satisfied; A pre-synchronization module, configured to: smoothly adjust the cylinder thrust of each partition before state switching by means of pre-synchronization.

[0019] The third aspect of the present invention provides a computer-readable storage medium, on which a computer program is stored. When the program is executed by a processor, it implements the steps in a method for determining the thrust of synchronous pushing and assembling cylinders of a shield tunneling machine as described above.

[0020] The fourth aspect of the present invention provides a computer device, including a computer-readable storage medium, a processor, and a computer program stored on the computer-readable storage medium and executable on the processor. When the processor executes the program, it implements the steps in a method for determining the thrust of synchronous pushing and assembling cylinders of a shield tunneling machine as described above.

[0021] Compared with the prior art, the beneficial effects of the present invention are as follows: The present invention adopts a segment-cylinder dynamic partitioning mechanism, which divides the cylinder partitions in real time according to the actual assembling requirements, so as to ensure the balance of the cylinder pressure in each partition under different assembling forms, and has strong dynamic adaptability, solving the problem that the traditional fixed partitioning is difficult to adapt to the diverse assembling requirements in the synchronous pushing and assembling state.

[0022] The present invention proposes a hybrid model combining CNN and Transformer, constructing two branches of historical observation data and planned cylinder thrust. The former extracts the influence of historical time steps (historical observation data from t - 49 to t) on the future attitude (at time t + 1) through CNN-Transformer, and the latter extracts the effect of the planned cylinder thrust (at time t + 1) on the future attitude (at time t + 1) through two fully connected layers. The two branches are connected by feature splicing and used to predict four tunneling attitude deviations. This model makes full use of the ability of CNN to process local time dependencies and the advantages of Transformer in long-term time series modeling, and can accurately predict the shield head horizontal deviation, shield head vertical deviation, shield tail horizontal deviation, and shield tail vertical deviation at time t + 1, providing high-quality model support for the calculation of the subsequent MPC cost function.

[0023] The present invention proposes a method for solving the thrust of each partition cylinder under the synchronous pushing and assembling state based on the model predictive control method (MPC). MPC is a dynamic control strategy that realizes the optimal tracking or regulation of the system by predicting the future system behavior and optimizing the control input in real time. Applying MPC to the control process of the shield synchronous pushing and assembling cylinder thrust can fully consider the influence of energy consumption and attitude deviation while ensuring force balance, and realize the coordinated closed-loop control of the shield machine synchronous pushing and assembling. MPC solves an online optimization problem in each sampling period, adaptively adjusts the control strategy according to the current system state and expected target, so as to realize the flexible switching between different cylinder thrust values. Adopting the rolling optimization and feedback correction strategies, continuously learning and optimizing according to the actual state of the shield machine to improve the prediction ability of the model.

[0024] In the process of feeding back the cylinder thrust decision result to the shield machine, the present invention introduces a pre-synchronization method, which not only realizes the smooth transition of the cylinder thrust during state switching, but also effectively reduces the mechanical impact and fluctuation risk during the system switching process, which is beneficial to improving the stability and safety of the shield machine propulsion process. Adopting a real-time closed-loop feedback mechanism to continuously monitor the state of the shield machine and dynamically correct the deviation during the pre-synchronization process to further suppress external disturbances and model errors, ensuring the stability and robustness of the overall control system.

[0025] Through the collaborative work of multiple modules such as data acquisition, preprocessing, real-time prediction, MPC decision-making, and pre-synchronization, the present invention constructs a set of robust and intelligent shield machine pushing and assembling synchronization control systems, which have strong adaptability and fault tolerance and are suitable for engineering applications in complex geological environments. BRIEF DESCRIPTION OF THE DRAWINGS

[0026] The specification drawings forming a part of the present invention are used to provide a further understanding of the present invention. The schematic embodiments of the present invention and their descriptions are used to explain the present invention and do not constitute an improper limitation to the present invention.

[0027] Figure 1 is a flowchart of a method for determining the thrust of a shield synchronous pushing and assembling cylinder in Embodiment 1 of the present invention; Figure 2 is a schematic diagram of the segment-cylinder dynamic partition in Embodiment 1 of the present invention; Figure 3 is a schematic diagram of the cylinder partition and the force on the cylinder in Embodiment 1 of the present invention; Figure 4 is a flowchart of the CNN-Transformer hybrid prediction model in Embodiment 1 of the present invention; Figure 5 is a schematic diagram of the intelligent decision-making of the cylinder thrust based on model predictive control in Embodiment 1 of the present invention; Figure 6It is a schematic structural diagram of a computer device according to Embodiment 4 of the present invention. Detailed implementation manners

[0028] To make the objectives, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention.

[0029] It should be noted that the following detailed descriptions are all illustrative and are intended to provide further descriptions of the present invention. Unless otherwise specified, all technical and scientific terms used herein have the same meanings as those commonly understood by those of ordinary skill in the technical field to which the present invention belongs.

[0030] Embodiment 1 This embodiment provides a method for determining the thrust of the shield synchronous pushing and assembling cylinders.

[0031] Using artificial intelligence methods to assist shield construction decision-making is an effective way to achieve scientific selection and regulation of construction parameters. A method for determining the thrust of the shield synchronous pushing and assembling cylinders provided in this embodiment divides the cylinder areas according to the segment position to ensure the balanced force of the unformed segments. Since the segment assembly method of staggered joints will cause changes in the segment position, the corresponding cylinder areas will also change accordingly. Therefore, an independent control system is set for each propulsion cylinder, a dynamic segment-cylinder zoning mechanism is established, and the cylinder areas are dynamically adjusted according to the real-time segment distribution form to achieve accurate response of the cylinders in each area under different segment assembly states; based on geological data and historical construction data, a hybrid prediction model based on CNN-Transformer is established to achieve accurate prediction of the shield head horizontal deviation, shield vertical deviation, shield tail horizontal deviation, and shield tail vertical deviation; the model predictive control method is adopted, and the total propulsion force, horizontal resultant moment, and vertical resultant moment are obtained based on the shield machine data acquisition system. The established tunneling attitude deviation prediction model is embedded in the cost function, and under the premise of ensuring the minimum cost and constraint conditions, the thrust distribution strategy of the cylinders in different areas under the synchronous pushing and assembling state is solved; finally, through the pre-synchronization method, the cylinder thrust is smoothly adjusted before the state switch, and the cylinder thrust result is fed back to the on-site shield construction through the pre-synchronization method, significantly reducing the system fluctuations and mechanical impacts caused by mutations, improving construction safety and equipment life, and realizing safe, efficient, and intelligent decision-making under the synchronous pushing and assembling state.

[0032] A method for determining the thrust of the shield synchronous pushing and assembling cylinders provided in this embodiment, as Figure 1 shown, includes the following steps: Step 1: Configure an independent control system for each propulsion cylinder and establish a segment-cylinder dynamic partitioning mechanism. Each segment corresponds to a cylinder partition. According to the segment distribution of the current segment ring to be assembled, the segment-cylinder partition is dynamically adjusted to achieve accurate and synchronous response of the cylinder pressure in each partition.

[0033] The key to the synchronous pushing and splicing technology is to keep the remaining cylinders producing the same resultant force and torque as the full cylinders in the pushing state while some cylinders are withdrawn, so as to minimize the posture deviation and ensure the safe and efficient construction of the shield machine along the designed axis. Figure 2 As shown in the figure, when the excavation length is greater than the width of one ring of segments, the shield machine enters the synchronous pushing and splicing state. When assembling a segment, the thrust cylinder of the corresponding partition needs to be withdrawn to leave enough space for the segment assembly, and the thrust of the remaining thrust cylinders is redistributed while ensuring that the resultant force and resultant moment remain unchanged.

[0034] In order to prevent the uncirculated segments from being damaged due to pressure eccentric loading, it is necessary to ensure that the same segment is evenly stressed. A ring of segments is divided into cylinder zones according to the number and position of the segments. Each segment corresponds to a zone, and the cylinder pressure inside each zone is the same. A ring of segments often includes multiple standard blocks, two adjacent blocks, and one capping block. The sizes of the three segments are significantly different, resulting in inconsistent numbers of cylinders corresponding to different segments-cylinder zones. The staggered assembly of segments results in different positions of the same segment in different rings, and there are differences in the thrust cylinder zones of different rings. Fixed zones are difficult to adapt to assembly requirements. It is necessary to establish a dynamic segment-thrust cylinder zone mechanism, demarcate the thrust cylinder zones in real time according to the segment assembly position of the current ring, and set an independent control system for each group of thrust cylinders to achieve accurate and rapid response of the cylinder pressure of each zone under different assembly forms.

[0035] The number of segments and cylinders is determined based on the actual tunnel diameter and shield machine diameter. Since the three segments (standard segment, adjacent segment, capping segment) have different sizes, the cylinder partition sizes corresponding to different segments are also different, so the number of cylinders included is different. Figure 3 For example, assume that a ring of shield tunnel segments consists of 6 segments, including 3 standard segments, 2 adjacent segments, and 1 capping segment. The propulsion system has a total of 20 evenly distributed hydraulic cylinders. Since one segment corresponds to one partition, A, B, C, D, F, and E are used to represent the six partitions. According to the design form of the tunnel and the segment, it is stipulated that the three zones A, B, and C each contain 4 propulsion cylinders, the two zones D and F each contain 3 propulsion cylinders, and E contains 2 propulsion cylinders.

[0036] by Figure 3 Take the six-zone propulsion system as an example. In order to avoid uneven force on the unlooped segments, the pressure of each propulsion cylinder in each zone is set to be the same. Denote the force on each propulsion cylinder in the $i$-th partition (where $i$ ranges from 1 to 6), and use to denote the propulsion cylinders included in the $i$-th partition, and use $k$ to denote the cylinder number (where $k$ ranges from 1 to 20, and $k = 1$ corresponds to the first cylinder in Area A).

[0037] Step 2: Collect the on-site construction data and geological data of the shield machine. After data preprocessing, establish a database for predicting the tunneling attitude deviation. Use the geological characteristics and the tunneling parameters of the shield machine as the input parameters of the tunneling attitude deviation prediction model to accurately predict the shield head horizontal deviation, shield head vertical deviation, shield tail horizontal deviation, and shield tail vertical deviation. Train the CNN-Transformer hybrid prediction model according to the current database, as Figure 4 shown, specifically including: (1) Data collection and preprocessing.

[0038] Collect the shield tunneling data and geological data, and preprocess the collected multivariate time series data, including normalization, outlier processing, and sliding window segmentation. Set the time window size to 50 time steps, that is, use the observation data of 50 consecutive historical time steps (from $t - 49$ to $t$) (covering the shield head horizontal deviation, shield head vertical deviation, shield tail horizontal deviation, shield tail vertical deviation, the actual thrust of each propulsion cylinder, propulsion speed, cutterhead torque, cutterhead speed, and formation type) as the input, and additionally introduce the planned pressure of each propulsion cylinder to be applied at $t + 1$ as the control instruction to jointly predict the attitude deviation at $t + 1$.

[0039] Normalize the original data, and use methods such as box plots and moving average filtering to process outliers and noise, and construct a database for regression prediction. The historical observation data is normalized independently by channel, and the planned cylinder pressure is normalized separately based on the mean and standard deviation of the historical cylinder pressure in the training set to avoid data leakage. Ensure that the planned quantity strictly corresponds to the $t + 1$ moment through timestamp verification.

[0040] (2) Model training and testing.

[0041] Constructing a tunneling attitude deviation prediction model based on the collaborative processing architecture of historical observation data and planned cylinder pressure: The historical observation data is multi-dimensional time series data for 50 consecutive time steps (from t-49 to the current time step t), including the shield head horizontal deviation, shield head vertical deviation, shield tail horizontal deviation, shield tail vertical deviation, actual thrust of each propulsion cylinder, propulsion speed, cutter head torque, cutter head rotation speed, and formation type. It is input into the CNN-Transformer joint encoder. The 1D convolutional layer extracts local fluctuation features, and the Transformer encoder receives the local fluctuation features extracted by the CNN and models the long-range dependencies between parameters through the self-attention mechanism, outputting a global state representation vector. The planning control branch independently processes the pressure commands of each propulsion cylinder at time step t+1, which are encoded into a high-dimensional operation intention vector through two fully connected layers, representing the impact of future operations on the tunneling attitude deviation. The two types of features are fused through feature concatenation and residual mapping: The global state vector and the operation intention vector are concatenated into a joint feature, which is non-linearly mapped through two fully connected layers (ReLU activation), and the output layer generates the predicted values of the shield head horizontal deviation, shield head vertical deviation, shield tail horizontal deviation, and shield tail vertical deviation at time step t+1.

[0042] Divide the training set and the test set according to 7:3, and train and test the established CNN-Transformer hybrid model. Finally, accurately predict the shield head horizontal deviation, shield vertical deviation, shield tail horizontal deviation, and shield tail vertical deviation at time step t+1, providing a basis for the subsequent calculation of the cylinder thrust.

[0043] Step 3: Adopt the model predictive control method to embed the tunneling construction attitude deviation prediction model established in Step 2 into the cost function calculation process. Under the condition of meeting the minimum cost and constraint conditions, calculate the thrust of each partition propulsion cylinder to ensure the safe and orderly progress of synchronous pushing and segment assembling, and realize the rapid and efficient construction of the shield according to the designed route.

[0044] (1) Obtain the shield head horizontal deviation, shield head vertical deviation, shield tail horizontal deviation, shield tail vertical deviation, actual thrust of each propulsion cylinder, propulsion speed, cutter head torque, cutter head rotation speed, total propulsion force , resultant moment along the x-axis , resultant moment along the y-axis through the shield construction data acquisition system. The obtained data is used in (2) the model predictive control method. Among them, the total propulsion force , resultant moment along the x-axis , resultant moment along the y-axis Known conditions in the prediction model; the horizontal deviation of the shield head, the vertical deviation of the shield head, the horizontal deviation of the shield tail, the vertical deviation of the shield tail, the actual thrust of each propulsion cylinder, the propulsion speed, the cutterhead torque, and the cutterhead rotation speed within a period of time provide data support for calculating the cost function (related to the tunneling attitude deviation prediction model).

[0045] Furthermore, when the shield machine switches from the pushing and segment - assembling separation state to the pushing and segment - assembling synchronous state, it is necessary to ensure that the total thrust and two resultant torques remain unchanged. Therefore, only at the moment of switching between the propulsion state and the pushing and segment - assembling synchronous state, the total propulsion force and the resultant torque along the x - axis and the resultant torque along the y - axis remain unchanged. The above three resultant forces and resultant torques are obtained through the shield tunneling data acquisition system. In addition, the construction speed of the shield machine is generally 10 - 40 mm / min, and the data acquisition frequency of the shield tunneling data acquisition system is 6 times / min. Therefore, the tunneling forward distance of the shield machine between two adjacent sampling times generally does not exceed 10 mm, and it can be considered that the geological conditions and construction conditions hardly change. Therefore, when the shield machine is in the synchronous pushing and segment - assembling state, it is also considered that the total propulsion force and the resultant torque along the x - axis and the resultant torque along the y - axis remain unchanged and are still obtained through the shield tunneling data acquisition system. In summary, in this embodiment, the total propulsion force and the resultant torque along the x - axis and the resultant torque along the y - axis take the same values as the previous moment and are all obtained through the shield tunneling data acquisition system.

[0046] (2) Adopt the model predictive control (MPC) framework, embed the tunneling attitude deviation prediction model established in step 2 into the cost function, and solve the sectional cylinder thrust distribution strategy that satisfies the minimum cost and meets the constraint conditions. The model predictive control framework mainly consists of five parts: a prediction model, a cost function, constraint conditions, rolling optimization, and feedback correction. As Figure 5 shown, the specific content is as follows: ① The prediction model is expressed as:

[0047]

[0048] As Figure 3 shown, r represents the radius of the hydraulic cylinder distribution circle, is the angle between the first cylinder in the first section (zone A) and the x - axis, is the angle between any two adjacent cylinders, represents the force magnitude of each propulsion cylinder in the i - th section, It represents the propulsion cylinder included in the i-th partition, and k represents the cylinder number.

[0049] ② Constraint conditions:

[0050]

[0051]

[0052]

[0053]

[0054] Among them, represents the cylinder thrust generated by each propulsion cylinder in the i-th partition, is the maximum thrust that a single propulsion cylinder can generate within the safe range, represent the horizontal deviations of the shield head and the shield tail respectively, represent the vertical deviations of the shield head and the shield tail respectively, represent the maximum allowable horizontal and vertical deviations during tunnel construction respectively.

[0055] ③ Cost function, constructing a two-layer cost function to compare the decision-making effects of all solutions in the feasible region: The first-layer cost function:

[0056] Among them, H represents the cost function of model predictive control, represent the weight coefficients of the energy consumption cost and the attitude deviation cost respectively.

[0057] The second-layer cost function: If there are multiple minimum solutions in the first-layer cost function at the same time, then by comparing the following formula, take the cylinder thrust strategy corresponding to the minimum value of the following formula:

[0058] Among them, is the weight coefficient, k is the cylinder number, is the maximum cylinder number (i.e., the total number of cylinders), represents the planned cylinder thrust of the k-th cylinder at time t + 1 (i.e., the feasible solution in the model predictive control framework), represents the actual cylinder thrust of the k-th cylinder at time t. The purpose of setting the second-layer cost function is to minimize the cylinder thrust change as much as possible and reduce the system fluctuation.

[0059] It should be noted that the four attitude deviations involved in ② and ③ are obtained through the tunneling attitude deviation prediction model established in Step 2. The specific calculation process is to input the thrusts of each cylinder planned in ① and historical data into the tunneling attitude deviation prediction model established in Step 2, and output the shield head horizontal deviation, shield vertical deviation, shield tail horizontal deviation, and shield tail vertical deviation that will be generated by this cylinder pressure distribution strategy. Then, judge whether the predicted four attitude deviations meet the requirements through the constraint conditions in ②. If they meet the requirements, calculate the corresponding cost value through the cost function in ③. If they do not meet the requirements of the constraint conditions in ②, directly delete the current solution. Traverse the feasible solution space and perform the above operations on each solution in turn.

[0060] ④Rolling optimization.

[0061] At each sampling moment, use the current system state to solve the optimal solution of the above cost function. The particle swarm optimization (PSO) algorithm is used to traverse all feasible solution spaces, and the solution that minimizes the cost function when meeting the constraint conditions is selected as the control input at the current moment. Since the system state and the external environment may change, MPC adopts a rolling horizon strategy, that is, re-optimize and calculate at each sampling moment to adapt to dynamic changes.

[0062] ⑤Feedback correction.

[0063] Real-time monitor the actual state of the shield machine (such as attitude deviation, propulsion speed, etc.), compare the actual measured value with the output of the prediction model, and calculate the error. According to the error, correct the two weights of the cost function in the MPC framework to improve the model accuracy. At the same time, adjust the control input to ensure that the system operates according to the desired trajectory.

[0064] Step 4: In order to achieve smooth switching of the cylinder thrust and reduce system fluctuations, a pre-synchronization method is introduced in the process of feeding back the cylinder thrust decision result obtained in Step 3 to the shield machine. Its core idea is to gradually adjust the current cylinder state to be close to the required state at the switching moment in advance before the execution of the state switching. The specific process is described as follows: (1) Prediction and determination of the switching moment.

[0065] Based on the rolling optimization and feedback correction mechanisms in the MPC framework, calculate and predict the change trend of the shield machine state in real time, so as to determine the accurate moment of the upcoming state switching. This prediction ensures that the controller can obtain the switching window in advance, providing a timing guarantee for the subsequent pre-synchronization.

[0066] (2) Pre-synchronization state adjustment.

[0067] Before the moment of switching, the thrust of the oil cylinder is smoothly adjusted through a pre-synchronization control strategy. The sigmoid function is used to generate a continuous and progressive control signal, enabling the thrust of the oil cylinder to gradually transition from the current state to the target state and avoiding sudden changes. The pre-synchronization adjustment ensures that the actual thrust of the oil cylinder has basically reached the expected switching state before switching.

[0068] (3) Real-time closed-loop feedback and correction.

[0069] During the pre-synchronization process, the system continuously collects the real-time operation data of the shield machine (such as the thrust of the oil cylinder, the total propulsion force, and the attitude deviation), and compares the actual state with the expected state. If there is a deviation, the controller will immediately adjust the control signal to ensure the smoothness and accuracy of the entire pre-synchronization process. This closed-loop feedback mechanism effectively suppresses the influence of external disturbances and model errors, ensuring the consistency of the system state at the switching moment.

[0070] Embodiment 2 This embodiment provides a system for determining the thrust of the synchronous pushing and assembling oil cylinders of a shield machine, which specifically includes: A zoning module, which is configured to: after the shield machine enters the synchronous pushing and assembling state, divide the segments of a ring of segments into oil cylinder zones according to the number and position of the segments, with each segment corresponding to one zone, and the oil cylinder pressure within each zone being the same; An oil cylinder thrust determination module, which is configured to: obtain the observed data at the historical moment and the oil cylinder pressures at the future moment, predict the attitude deviation at the future moment through a tunneling attitude deviation prediction model; obtain the total propulsion force, the resultant moment in the horizontal direction, and the resultant moment in the vertical direction, and combine the attitude deviation at the future moment to solve for the thrust of each zone of the oil cylinder when meeting the minimum energy consumption cost and attitude deviation cost through a model predictive control framework; A pre-synchronization module, which is configured to: smoothly adjust the thrust of each zone of the oil cylinder before the state switch through a pre-synchronization method.

[0071] It should be noted here that each module in this embodiment corresponds one by one to each step in Embodiment 1, and the specific implementation process is the same, so it will not be repeated here.

[0072] Embodiment 3 This embodiment provides a computer-readable storage medium, on which a computer program is stored, and when the program is executed by a processor, it implements the steps in a method for determining the thrust of the synchronous pushing and assembling oil cylinders of a shield machine as described in Embodiment 1 above.

[0073] Embodiment 4 This embodiment provides a computer device, such as Figure 6As shown, it includes a computer-readable storage medium 1003, a processor 1001, a communication interface 1002, and a computer program stored on the computer-readable storage medium 1003 and executable on the processor 1001. Among them, the processor 1001, the communication interface 1002, and the computer-readable storage medium 1003 can be connected through a bus or other means. Among them, the communication interface 1002 is used to receive and send data, and when the processor 1001 executes the program, it implements the steps in a method for determining the thrust of a shield synchronous pushing and assembling oil cylinder as described in Embodiment 1 above.

[0074] The above are only the preferred embodiments of the present invention and are not used to limit the present invention. For those skilled in the art, the present invention can have various changes and modifications. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

Claims

1. A method for determining the thrust of a shield synchronous pushing and assembling oil cylinder, characterized in that, Including: After the shield machine enters the synchronous pushing and segment assembling state, divide the oil cylinder zones according to the number and position of segments for one ring. Each segment corresponds to one zone, and the oil cylinder pressure inside each zone is the same. Obtain the observed data at the historical moment and the oil cylinder pressures at the future moment. Through the tunneling attitude deviation prediction model, predict the attitude deviation at the future moment. Obtain the total propulsion force, the resultant moment in the horizontal direction, and the resultant moment in the vertical direction. Combine with the attitude deviation at the future moment, and through the model predictive control framework, solve the oil cylinder thrusts of each zone when meeting the minimum energy consumption cost and attitude deviation cost. Through the pre-synchronization method, smoothly adjust the oil cylinder thrusts of each zone before the state switch.

2. The method for determining the thrust of the shield synchronous pushing and assembling oil cylinder according to claim 1, wherein The tunneling attitude deviation prediction model includes two branches: Branch one takes the observed data at the historical moment as the input, and uses the CNN-Transformer joint encoder to obtain the global state representation vector; Branch two takes the oil cylinder pressures at the future moment as the input, and uses a two-layer fully connected network for encoding to obtain the operation intention vector. After the global state representation vector and the operation intention vector are fused through feature splicing and residual mapping, and then non-linearly mapped through a two-layer fully connected network, the attitude deviation at the future moment is obtained.

3. The shield tunneling attitude deviation prediction model according to claim 1, characterized in that For the tunneling attitude deviation prediction model, the observed data at the historical moment includes the shield head horizontal deviation, the shield head vertical deviation, the shield tail horizontal deviation, the shield tail vertical deviation, the actual thrust of each propulsion oil cylinder, the propulsion speed, the cutterhead torque, the cutterhead rotation speed, and the formation type. The attitude deviation at the future moment includes the shield head horizontal deviation, the shield head vertical deviation, the shield tail horizontal deviation, and the shield tail vertical deviation.

4. The method for determining the thrust of the shield synchronous pushing and assembling oil cylinder according to claim 1, wherein, During the solving process of the oil cylinder thrusts of each zone, the oil cylinder thrusts of each zone and the attitude deviation are constrained.

5. The method for determining the thrust of the shield synchronous pushing and assembling oil cylinder according to claim 1, characterized in that The prediction model adopted during the solving process of the oil cylinder thrusts of each zone is expressed as: where r represents the radius of the distribution circle of the hydraulic cylinders, is the angle between the first cylinder in the first partition (zone A) and the x-axis, is the angle between any two adjacent cylinders, represents the force on each propulsion cylinder in the i-th partition, represents the propulsion cylinders included in the i-th partition, and k represents the cylinder number.

6. The method for determining the thrust of the shield synchronous pushing and assembling oil cylinder according to claim 1, wherein, During the solving process of the oil cylinder thrusts of each zone, a rolling time domain strategy is adopted, and the particle swarm optimization algorithm is used to traverse all feasible solution spaces.

7. A method for determining the thrust of a shield synchronous pushing and assembling oil cylinder as claimed in claim 1, characterized in that, During the solving process of the oil cylinder thrusts of each zone, the actual state of the shield machine is monitored in real time. Compare the actual measured value with the output of the prediction model, calculate the error, and correct the weight of the cost function of the model predictive control framework according to the error.

8. A thrust determination system for shield synchronous pushing and assembling cylinders, characterized in that, Including: A zoning module, which is configured to: after the shield machine enters the synchronous pushing and segment assembling state, divide the oil cylinder zones according to the number and position of segments for one ring. Each segment corresponds to one zone, and the oil cylinder pressure inside each zone is the same. An oil cylinder thrust determination module, which is configured to: obtain the observed data at the historical moment and the oil cylinder pressures at the future moment. Through the tunneling attitude deviation prediction model, predict the attitude deviation at the future moment. Obtain the total propulsion force, the resultant moment in the horizontal direction, and the resultant moment in the vertical direction. Combine with the attitude deviation at the future moment, and through the model predictive control framework, solve the oil cylinder thrusts of each zone when meeting the minimum energy consumption cost and attitude deviation cost. A pre-synchronization module, which is configured to: through the pre-synchronization method, smoothly adjust the oil cylinder thrusts of each zone before the state switch.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the program is executed by a processor, it implements the steps in a method for determining the oil cylinder thrust of shield synchronous pushing and segment assembling as described in any one of claims 1-7.

10. A computer device, comprising a computer-readable storage medium, a processor, and a computer program stored on the computer-readable storage medium and executable on the processor, characterized in that, When the processor executes the program, it implements the steps in a method for determining the thrust of a shield synchronous pushing and assembling oil cylinder as described in any one of claims 1 to 7.

Images