A large model driven hydraulic synchronous lifting control system and method

Abstract

The application discloses a large model driven hydraulic synchronous lifting control system and method, and relates to the technical field of hydraulic synchronous lifting control.The control system comprises a hydraulic lifting execution cluster, a sensor network, an edge calculation and control node, a large model intelligent decision center and a man-machine interaction interface; a core decision large model is pre-trained by combining a hydraulic synchronous lifting physical mechanism and a large amount of operation data, the large model intelligent decision center and the edge calculation node constitute a collaborative architecture of edge execution and center decision, and working condition recognition, dynamic prediction and control strategy optimization are realized.The control method completes a lifting task through a closed loop process of sensor data collection, edge node bottom layer control, large model intelligent decision optimization and instruction execution.The application can improve multi-hoisting point synchronous precision, predict risks in advance, realize man-machine collaborative management and control, has efficient response and strong robustness, and is suitable for various large structure lifting scenes.

Description

Technical Field

[0001] This invention relates to the field of hydraulic synchronous lifting control technology, specifically to a large-scale hydraulic synchronous lifting control system and method. Background Technology

[0002] Hydraulic synchronous lifting technology is an advanced construction technique widely used for the overall lifting of large structures (such as steel structures, bridges, and roofs). Its core requirement is to ensure the stability of the lifted components and the precision of the lifting through multi-point coordinated control, avoiding safety accidents such as uneven stress, deformation, or even falling of components due to synchronous deviation of the lifting points.

[0003] Traditional hydraulic synchronous lifting systems typically combine computer control, electro-hydraulic proportional valves, sensor detection, and PID control algorithms to achieve synchronous lifting, load balancing, and attitude control across multiple lifting points. This traditional control system suffers from limitations such as strong parameter dependence, insufficient adaptive capability, weak predictive ability, outdated monitoring methods, and low levels of information technology. In recent years, large-scale AI models (such as Transformer, GPT series, and reinforcement learning models) have demonstrated significant potential in industrial control, enabling adaptive decision-making and predictive control through data-driven approaches.

[0004] The purpose of this invention is to overcome the shortcomings of existing synchronous lifting technologies and provide a large-model driven hydraulic synchronous lifting control system and method. By introducing the intelligent perception, prediction and decision-making capabilities of the large model, the control accuracy, adaptability, safety and intelligence level of the system under load conditions are improved. Summary of the Invention

[0005] To address the technical shortcomings of existing hydraulic synchronous lifting control systems, such as poor adaptability, weak predictive ability, and insufficient control level, this invention provides a large-model-driven hydraulic synchronous lifting control system and method. By introducing a core decision-making large model pre-trained based on the physical mechanism of hydraulic synchronous lifting and massive operational data, the system's working condition identification, dynamic prediction, and intelligent decision-making capabilities are enhanced, enabling precise control, risk prediction, and safety management of the lifting process of large structures.

[0006] To achieve the above objectives, the present invention provides the following technical solution:

[0007] On the one hand, the present invention provides a large-scale model-driven hydraulic synchronous lifting control system, including a hydraulic lifting execution cluster, a sensor network, edge computing and control nodes, a large-scale model intelligent decision-making center, and a human-machine interface;

[0008] The hydraulic lifting execution cluster includes a cluster-type lifting device, a load-bearing system, a hydraulic pump station, and a valve group, which is used to receive control signals from edge computing and control nodes and execute lifting operations on the lifted components.

[0009] The sensor network is deployed on the lifted component and the hydraulic lifting actuator cluster to collect real-time data on the position of the lifting device, the status of the anchor, the oil pressure of the main lifting system, and the synchronous height difference signal of each lifting point, and transmits the collected real-time data to the edge computing and control node.

[0010] The edge computing and control nodes are respectively connected to the hydraulic lifting execution cluster and the sensor network to receive and preliminarily process the real-time data, and execute the underlying closed-loop control based on the preliminarily processed real-time data with multi-lifting point synchronous error compensation as the core to maintain the basic operation of the system. At the same time, they receive and execute the optimized control instructions issued by the large model intelligent decision-making center.

[0011] The large-scale intelligent decision-making center is communicatively connected to the edge computing and control nodes. It receives the pre-processed real-time data uploaded by the edge computing and control nodes and combines it with the system's historical operating data. Through the core decision-making large-scale model, it performs working condition identification, dynamic prediction, and control strategy and parameter optimization, and issues optimized control commands to the edge computing and control nodes. The core decision-making large-scale model is obtained by pre-training based on the physical mechanism of hydraulic synchronous lifting and massive operating data.

[0012] The human-computer interaction interface is communicatively connected to the large model intelligent decision-making center and the edge computing and control node, respectively. It is used to visualize the system status, dynamic prediction results, control decisions and their interpretation information, and to receive intervention instructions from operators and feed them back to the large model intelligent decision-making center and / or the edge computing and control node.

[0013] Preferably, the large-scale intelligent decision-making center includes a multimodal data fusion module, a digital twin module, a core decision-making large model, and a strategy output and interpretation module;

[0014] The multimodal data fusion module is used to align, clean, reduce noise, extract features, and construct state vectors for the pre-processed real-time data and the system's historical operating data, and output a standardized system state vector.

[0015] The digital twin module constructs a high-fidelity virtual model based on the system state vector and the physical parameters of the hydraulic lifting system, realizing real-time synchronization, state mapping and simulation between the physical system and the virtual model.

[0016] The core decision-making model is used to receive simulation data from the system state vector and the digital twin module, and to perform operating condition identification, dynamic prediction, and control strategy and parameter optimization.

[0017] The strategy output and interpretation module is used to convert the output of the core decision model into executable control instructions, generate corresponding decision interpretation information, and perform permission verification and security verification on the control instructions.

[0018] Preferably, the core decision-making big model is built and trained based on the Transformer architecture.

[0019] Preferably, the training data of the core decision-making big model integrates hydraulic system dynamic model simulation data, historical lifting project operation data, fault case data, and expert operation rules. Its training objective is to minimize the synchronization error of each lifting point and maximize the improvement efficiency and system safety margin.

[0020] Preferably, the edge computing and control node and the large model intelligent decision-making center constitute a collaborative architecture of edge execution and central decision-making; wherein, the edge computing and control node is responsible for high real-time closed-loop control, and the large model intelligent decision-making center is responsible for non-real-time decision-making and predictive optimization.

[0021] Preferably, the valve assembly includes an electro-hydraulic proportional valve, and the sensor network includes a laser rangefinder, a position switch, a limit switch, and a hydraulic pressure sensor; wherein:

[0022] The laser rangefinder is deployed at each suspension point to collect the synchronous height difference between the suspension points;

[0023] The position switches are installed at the actuation ends of the upper and lower anchorages to detect the status of the anchorages;

[0024] The limit switches are deployed at multiple points on the outside of the main hydraulic cylinder of the elevator to collect the current position of the piston rod of the hydraulic cylinder of the elevator.

[0025] The oil pressure sensor is installed at the outlet of the electro-hydraulic proportional valve and the oil inlet of the main cylinder large chamber to collect the oil pressure of the lifting main system.

[0026] Preferably, the edge computing and control node adopts an industrial PLC and an embedded system, interacts with the hydraulic lifting actuator cluster and sensor network via a CAN bus, and runs a PID-based synchronous control algorithm to maintain the basic operation of the system.

[0027] On the other hand, the present invention also provides a large-model driven hydraulic synchronous lifting control method, applied to the above-mentioned large-model driven hydraulic synchronous lifting control system, the method comprising the following steps:

[0028] S1: The real-time data of the hoisting system is continuously collected through the sensor network. The real-time data includes the current position of the piston rod of the hoisting cylinder, the status of the anchor, the oil pressure of the main hoisting system, and the synchronous height difference signal of each hoisting point.

[0029] S2: The edge computing and control node receives and preliminarily processes the real-time data collected by the sensor network. Based on the preliminarily processed real-time data, it executes the underlying closed-loop control with multi-suspension point synchronization error compensation as the core to maintain the basic operation of the system. At the same time, it uploads the preliminarily processed real-time data to the large model intelligent decision-making center.

[0030] S3: After receiving the pre-processed real-time data, the large-scale intelligent decision-making center combines the system's historical operating data and uses the core decision-making large model to perform condition identification, dynamic prediction, and control strategy and parameter optimization to generate executable control commands and corresponding decision interpretation information. Subsequently, the control commands are sent to the edge computing and control nodes, and the decision interpretation information is pushed to the human-computer interaction interface.

[0031] S4: The edge computing and control node adjusts its control behavior according to the received control commands;

[0032] S5: Repeat steps S1 to S4 until the lifting and lowering task of the lifted component is completed.

[0033] Preferably, in step S3, the dynamic prediction includes the prediction of the synchronous deviation of each lifting point, the stress distribution of key points of the lifted component, and the trend of hydraulic pressure change in the main lifting system within a future preset time period.

[0034] Preferably, in step S3, the control strategy and parameter optimization includes generating a globally optimal control strategy and dynamically adjusting the PID controller parameters.

[0035] Compared with the prior art, the present invention has at least one of the following advantages or beneficial effects:

[0036] 1) Combining precise synchronous control with strong robustness: The core decision-making model is pre-trained based on the physical mechanism of hydraulic synchronous lifting and massive operational data. It can accurately identify complex working conditions such as wind load interference and automatically optimize control strategies and parameters. It can effectively reduce the synchronous deviation of multiple lifting points and ensure the stability of the lifting process. It can also intelligently compensate for the dynamic changes of the system and the influence of external disturbances, adapting to the lifting needs of large structures of different tonnages and different scenarios, and has a wider range of applications.

[0037] 2) Enhanced prediction and risk management capabilities: By using a high-fidelity virtual model built with a digital twin module for simulation and deduction, combined with the prediction function of the core decision-making big model, potential risks such as excessive synchronous deviation of lifting points, stress concentration of components, and hydraulic system failures can be detected in advance. Then, through the fault diagnosis function, early warnings can be issued in a timely manner and disposal solutions can be provided, which greatly improves construction safety.

[0038] 3) More convenient human-machine collaborative management: The human-machine interface can clearly display the system's operating status, decision-making basis, and early warning information. Operators can intuitively understand the control logic and participate in management through intervention commands, taking into account both the autonomy of automatic control and the reliability of manual operation.

[0039] 4) The system has a reasonable architecture and high efficiency: It adopts a collaborative architecture that combines edge execution and central decision-making. Edge computing and control nodes are responsible for high real-time control to ensure rapid response of actions; the large model intelligent decision-making center is responsible for complex calculations to ensure accurate decision-making and avoid the response delay or rough decision-making problems that may occur in a single architecture.

[0040] 5) Improved intelligence and construction efficiency: The system can autonomously complete control and optimization decisions, reduce manual intervention, and improve construction efficiency and project quality. At the same time, the large model can continuously accumulate experience from new engineering data, form reusable construction knowledge and optimal operation solutions, and provide technical support for subsequent similar improvement projects. Attached Figure Description

[0041] The invention, its features, shape, and advantages will become more apparent from the following detailed description of non-limiting embodiments with reference to the accompanying drawings. Like reference numerals denote like parts throughout the drawings. The drawings are not drawn to scale; their focus is on illustrating the gist of the invention.

[0042] Figure 1 This is a schematic diagram of the overall architecture of the large-model driven hydraulic synchronous lifting control system in an embodiment of the present invention.

[0043] Figure 2 This is a diagram showing the internal functional modules of the large-scale intelligent decision-making center in an embodiment of the present invention.

[0044] Figure 3 This is a flowchart of the control method for a large-model-driven hydraulic synchronous lifting control system in an embodiment of the present invention. Detailed Implementation

[0045] The present invention will be further described below with reference to the accompanying drawings and specific embodiments, but these are not intended to limit the scope of the invention.

[0046] Example 1:

[0047] like Figure 1 As shown, this embodiment discloses a large-model-driven hydraulic synchronous lifting control system, including a hydraulic lifting execution cluster, a sensor network, edge computing and control nodes, a large-model intelligent decision-making center, and a human-machine interface. The connection relationships and functions of each component are as follows:

[0048] The hydraulic lifting execution cluster includes a cluster of lifting devices (specifically, multiple 350t lifting cylinders), a load-bearing system, a hydraulic pump station, and a valve group. It is used to receive control signals from edge computing and control nodes and execute the lifting and lowering operations (i.e., ascending and descending operations) of the lifted components. The hydraulic pump station provides hydraulic power to the system, and the valve group uses electro-hydraulic proportional valves to achieve precise speed control of each lifting device. The lifting devices work in conjunction with the load-bearing system through the extension and retraction of the cylinders to complete the stable lifting and lowering of the lifted components.

[0049] A sensor network is deployed within the lifted component and the hydraulic lifting actuator cluster to collect key status data in real time during the lifting process. This includes the current position of the lifting cylinder piston rod, the anchor tension status, the hydraulic pressure of the main lifting system, and the synchronous height difference signal at each lifting point. The collected real-time data (i.e., raw data) is transmitted to the edge computing and control node. Specifically, the sensor network includes a laser rangefinder, position switches, limit switches, and hydraulic pressure sensors. The laser rangefinder is deployed at each lifting point to collect the synchronous height difference. The position switches are deployed at the upper and lower anchor actuator ends to detect the anchor tension status. Multiple limit switches are equidistantly deployed at multiple points on the outside of the main lifting cylinder to collect the current position of the lifting cylinder piston rod. The hydraulic pressure sensors are deployed at the outlet of the electro-hydraulic proportional valve and the oil inlet of the main cylinder (i.e., the oil inlet connected to the main cylinder's large chamber) to collect the hydraulic pressure of the main lifting system.

[0050] The aforementioned edge computing and control nodes are connected to the hydraulic lifting execution cluster and sensor network via CAN bus, respectively. They are built using high-performance industrial PLCs and embedded systems. Their core functions include: first, receiving and initially processing real-time data collected by the sensor network (specifically, filtering and extracting the real-time data); second, executing low-level closed-loop control based on multi-point synchronous error compensation based on the initially processed real-time data, and running basic PID synchronous control algorithms to maintain basic system operation; and third, receiving and executing optimized control commands issued by the large-scale intelligent decision-making center.

[0051] The aforementioned large-scale intelligent decision-making center, deployed on a local high-performance server or private cloud, is the core of the system's intelligent decision-making. This large-scale intelligent decision-making center communicates with edge computing and control nodes to receive pre-processed real-time data uploaded by the edge computing and control nodes (including the height difference of each lifting point, the main system oil pressure, the current position of the piston rod of the lifting cylinder, etc.). It combines this data with the system's historical operating data stored in the local high-performance server or private cloud (specifically, the operating parameters, working conditions, and fault records related to the hydraulic synchronous lifting process collected and stored during the system's historical operating cycle). Through the pre-trained core decision-making large-scale model, it performs working condition identification, dynamic prediction, decision optimization, and control strategy and parameter optimization, and issues optimized control commands to the edge computing and control nodes. This core decision-making large-scale model is obtained through pre-training based on the physical mechanism of hydraulic synchronous lifting and massive operating data.

[0052] The aforementioned human-computer interaction interface is communicatively connected to the large-scale intelligent decision-making center and the edge computing and control node, respectively. It is used to visualize and display the system status, dynamic prediction results, control decisions, and explanatory information. At the same time, it receives intervention commands from operators (such as emergency stop commands, parameter fine-tuning commands, etc.) and feeds them back to the large-scale intelligent decision-making center and / or the edge computing and control node. (Specifically, if several pre-commands are only emergency stop commands, they are fed back to the edge computing and control node; if several pre-commands are parameter fine-tuning commands, they are fed back to the large-scale intelligent decision-making center; if several pre-commands include both emergency stop commands and parameter fine-tuning commands, the emergency stop command is fed back to the edge computing and control node, and the parameter fine-tuning command is fed back to the large-scale intelligent decision-making center.) This enables human-computer collaborative management and control.

[0053] like Figure 2As shown, the aforementioned large-scale intelligent decision-making center consists of four core modules: a multimodal data fusion module, a digital twin module, a core decision-making large model, and a strategy output and interpretation module. The multimodal data fusion module receives pre-processed real-time data uploaded from edge computing and control nodes. Its function is to sequentially align, clean, denoise, extract features, and construct state vectors from the pre-processed real-time data and historical system operation data stored on local high-performance servers or in private clouds. The output is a high-quality, standardized system state vector for use by downstream modules. The digital twin module receives the real-time system state vector from the data fusion module and the physical parameters of the hydraulic lifting system (such as component material properties and hydraulic system performance parameters). Based on the system state vector and the physical parameters of the hydraulic lifting system, it constructs a high-fidelity virtual model, achieving real-time synchronization, state mapping, and simulation between the physical system and the virtual model. This module's functions include high-fidelity modeling, real-time synchronization and state mapping, and simulation, outputting real-time system status display and prediction data for future dynamics (i.e., simulation data). The core decision-making model module takes as input the state vector output from the data fusion module and the simulation data output from the digital twin module. Its core functions include: 1. Operating condition identification: specifically identifying the current load distribution, structural stiffness changes, and external wind load conditions; 2. Multi-step dynamic prediction: accurately predicting the synchronous deviation of each suspension point and the evolution of structural stress in future periods, and simultaneously completing fault diagnosis, location, and equipment health status determination; 3. Control strategy and parameter optimization: covering parameter adaptive tuning, path dynamic optimization, and intelligent decision generation, specifically including generating advanced control strategies or dynamically optimizing PID controller parameters, and dynamically planning the optimal lifting speed and path. The output of this core decision-making model module is the original, high-level decision command. The input to the control strategy and interpretation module is the original decision command generated by the core decision model module. The functions of this module are command conversion, interpretability analysis, and permission and security verification. Specifically, the control strategy and interpretation module converts the original decision command output by the core decision model into executable control commands, generates explanatory information on the decision basis (including working condition identification basis, dynamic prediction logic, decision optimization reasons and fault diagnosis descriptions), and performs permission verification and security verification on the control commands. The output is the optimized control command to the edge control node and the decision interpretation information to the human-machine interface.

[0054] Preferably, the aforementioned core decision-making model is trained based on the Transformer architecture. Furthermore, the training data for this core decision-making model integrates simulation data from hydraulic system dynamics models generated using software such as AMESim, historical lifting project operation data (i.e., sensor data, control commands, and equipment parameters from historical lifting projects), fault case data (i.e., fault records), and expert operating rules. Its training objective is to minimize the synchronization error of each lifting point and maximize lifting efficiency and system safety margin. By integrating physical mechanisms, it ensures that the model's decisions conform to the engineering laws of hydraulic lifting, and by training with massive amounts of operational data, it enhances the model's adaptability to complex working conditions.

[0055] The edge computing and control node of this invention and the large model intelligent decision-making center constitute a collaborative architecture of edge execution and central decision-making (the central decision-making includes cloud decision-making and local decision-making); the edge computing and control node is responsible for high real-time closed-loop control to ensure rapid response of the improved actions; the large model intelligent decision-making center is responsible for non-real-time decision-making and predictive optimization decision-making, avoiding complex calculations from occupying edge node resources and ensuring the real-time performance and stability of system control.

[0056] Example 2:

[0057] like Figure 3 As shown, this embodiment provides a large-model-driven hydraulic synchronous lifting control method. This method is applied to the aforementioned large-model-driven hydraulic synchronous lifting control system. Specifically, the method includes the following steps:

[0058] S1: The sensor network continuously collects real-time data of the hoisting system, including the current position of the piston rod of the hoisting cylinder, the status of the anchor, the hydraulic pressure of the main hoisting system, and the synchronous height difference signal of each hoisting point;

[0059] S2: The edge computing and control node receives and preliminarily processes the real-time data collected by the sensor network. Based on the preliminarily processed real-time data, it executes the underlying closed-loop control (i.e., the basic synchronization control algorithm) with multi-suspension point synchronization error compensation as the core to maintain the basic operation of the system. At the same time, it uploads the preliminarily processed real-time data to the large model intelligent decision-making center.

[0060] S3: After receiving the pre-processed real-time data, the large-scale intelligent decision-making center combines it with historical system operation data. Through the core decision-making large-scale model, it performs condition identification, dynamic prediction, and control strategy and parameter optimization to generate executable control commands and corresponding decision interpretation information. Subsequently, the control commands are sent to the edge computing and control nodes, and the decision interpretation information is pushed to the human-machine interface. Specifically, the dynamic prediction includes prediction of the synchronization deviation of each lifting point, stress distribution of key points of the lifted components, and the trend of hydraulic pressure change in the main lifting system within a preset time period, as well as fault diagnosis, location, and equipment health status determination. The control strategy and parameter optimization includes generating advanced control strategies and dynamically optimizing PID controller parameters, such as a globally optimal control strategy.

[0061] In step S3, when it is predicted that the synchronization deviation will exceed the threshold or a potential fault risk is identified, the large model intelligent decision center will generate early warning information and correction schemes, and provide them to the operator through the human-computer interaction interface.

[0062] S4: Edge computing and control nodes adjust their control behavior based on received control commands.

[0063] S5: Repeat steps S1 to S4 until the lifting and lowering task of the lifted component is completed.

[0064] It is not difficult to see that this embodiment is a method embodiment corresponding to the above-described embodiment of the large-model-driven hydraulic synchronous lifting control system. This embodiment can be implemented in conjunction with the above-described embodiment of the large-model-driven hydraulic synchronous lifting control system. The relevant technical details mentioned in the above-described embodiment of the large-model-driven hydraulic synchronous lifting control system are still valid in this embodiment, and will not be repeated here to reduce repetition. Accordingly, the relevant technical details mentioned in this embodiment can also be applied to the above-described embodiment of the large-model-driven hydraulic synchronous lifting control system.

[0065] Those skilled in the art should understand that variations can be implemented by combining existing technology with the above embodiments, which will not be elaborated here. Such variations do not affect the essence of the present invention, and will not be elaborated here either.

[0066] The preferred embodiments of the present invention have been described above. It should be understood that the present invention is not limited to the specific embodiments described above, and the devices and structures not described in detail should be understood as being implemented in a conventional manner in the art. Any person skilled in the art can make many possible variations and modifications to the technical solutions of the present invention using the methods and techniques disclosed above, or modify them into equivalent embodiments with equivalent changes, without departing from the scope of the present invention. This does not affect the essential content of the present invention. Therefore, any simple modifications, equivalent changes, and modifications made to the above embodiments based on the technical essence of the present invention without departing from the content of the present invention's technical solutions still fall within the protection scope of the present invention.

Claims

1. A large-scale model-driven hydraulic synchronous lifting control system, characterized in that, This includes a hydraulic lifting actuator cluster, a sensor network, edge computing and control nodes, a large-scale intelligent decision-making center, and a human-machine interface; The hydraulic lifting execution cluster includes a cluster-type lifting device, a load-bearing system, a hydraulic pump station, and a valve group, which is used to receive control signals from edge computing and control nodes and execute lifting operations on the lifted components. The sensor network is deployed on the lifted component and the hydraulic lifting actuator cluster to collect real-time data on the position of the lifting device, the status of the anchor, the oil pressure of the main lifting system, and the synchronous height difference signal of each lifting point, and transmits the collected real-time data to the edge computing and control node. The edge computing and control nodes are respectively connected to the hydraulic lifting execution cluster and the sensor network to receive and preliminarily process the real-time data, and execute the underlying closed-loop control based on the preliminarily processed real-time data with multi-lifting point synchronous error compensation as the core to maintain the basic operation of the system. At the same time, they receive and execute the optimized control instructions issued by the large model intelligent decision-making center. The large-scale intelligent decision-making center is connected to the edge computing and control nodes to receive the pre-processed real-time data uploaded by the edge computing and control nodes. Combined with the system's historical operating data, the core decision-making large model performs working condition identification, dynamic prediction, and control strategy and parameter optimization, and issues optimized control commands to the edge computing and control nodes. The core decision-making model is obtained by pre-training based on the physical mechanism of hydraulic synchronous lifting and massive operational data. The human-computer interaction interface is communicatively connected to the large model intelligent decision-making center and the edge computing and control node, respectively. It is used to visualize the system status, dynamic prediction results, control decisions and their interpretation information, and to receive intervention instructions from operators and feed them back to the large model intelligent decision-making center and / or the edge computing and control node.

2. The large-model driven hydraulic synchronous lifting control system according to claim 1, characterized in that, The large-scale intelligent decision-making center includes a multimodal data fusion module, a digital twin module, a core decision-making large model, and a strategy output and interpretation module; The multimodal data fusion module is used to align, clean, reduce noise, extract features, and construct state vectors for the pre-processed real-time data and the system's historical operating data, and output a standardized system state vector. The digital twin module constructs a high-fidelity virtual model based on the system state vector and the physical parameters of the hydraulic lifting system, realizing real-time synchronization, state mapping and simulation between the physical system and the virtual model. The core decision-making model is used to receive simulation data from the system state vector and the digital twin module, and to perform operating condition identification, dynamic prediction, and control strategy and parameter optimization. The strategy output and interpretation module is used to convert the output of the core decision model into executable control instructions, generate corresponding decision interpretation information, and perform permission verification and security verification on the control instructions.

3. The large-model driven hydraulic synchronous lifting control system according to claim 1 or 2, characterized in that, The core decision-making model is built and trained based on the Transformer architecture.

4. The large-model driven hydraulic synchronous lifting control system according to claim 1 or 2, characterized in that, The training data of the core decision-making big model integrates hydraulic system dynamic model simulation data, historical lifting project operation data, fault case data, and expert operation rules. Its training objective is to minimize the synchronization error of each lifting point and maximize the improvement efficiency and system safety margin.

5. The large-model driven hydraulic synchronous lifting control system according to claim 1, characterized in that, The edge computing and control nodes and the large model intelligent decision-making center constitute a collaborative architecture of edge execution and central decision-making; wherein, the edge computing and control nodes are responsible for high real-time closed-loop control, and the large model intelligent decision-making center is responsible for non-real-time decision-making and predictive optimization.

6. The large-model driven hydraulic synchronous lifting control system according to claim 1, characterized in that, The valve assembly includes an electro-hydraulic proportional valve, and the sensor network includes a laser rangefinder, a position switch, a limit switch, and a hydraulic pressure sensor; wherein: The laser rangefinder is deployed at each suspension point to collect the synchronous height difference between the suspension points; The position switches are installed at the actuation ends of the upper and lower anchorages to detect the status of the anchorages; The limit switches are deployed at multiple points on the outside of the main hydraulic cylinder of the elevator to collect the current position of the piston rod of the hydraulic cylinder of the elevator. The oil pressure sensor is installed at the outlet of the electro-hydraulic proportional valve and the oil inlet of the main cylinder large chamber to collect the oil pressure of the lifting main system.

7. The large-model driven hydraulic synchronous lifting control system according to claim 1, characterized in that, The edge computing and control node adopts an industrial PLC and embedded system, interacts with the hydraulic lifting actuator cluster and sensor network via CAN bus, and runs a PID-based synchronous control algorithm to maintain the basic operation of the system.

8. A hydraulic synchronous lifting control method driven by a large model, characterized in that, A hydraulic synchronous lifting control system for large-scale model drive, applicable to any one of claims 1 to 7, the method comprising the following steps: S1: The real-time data of the hoisting system is continuously collected through the sensor network. The real-time data includes the current position of the piston rod of the hoisting cylinder, the status of the anchor, the oil pressure of the main hoisting system, and the synchronous height difference signal of each hoisting point. S2: The edge computing and control node receives and preliminarily processes the real-time data collected by the sensor network. Based on the preliminarily processed real-time data, it executes the underlying closed-loop control with multi-suspension point synchronization error compensation as the core to maintain the basic operation of the system. At the same time, it uploads the preliminarily processed real-time data to the large model intelligent decision-making center. S3: After receiving the pre-processed real-time data, the large-scale intelligent decision-making center combines it with the system's historical operating data and uses the core decision-making large-scale model to perform condition identification, dynamic prediction, and control strategy and parameter optimization to generate executable control commands and corresponding decision interpretation information. Subsequently, the control commands are sent to the edge computing and control nodes, and the decision interpretation information is pushed to the human-computer interaction interface. S4: The edge computing and control node adjusts its control behavior according to the received control commands; S5: Repeat steps S1 to S4 until the lifting and lowering task of the lifted component is completed.

9. The hydraulic synchronous lifting control method driven by a large model according to claim 8, characterized in that, In step S3, the dynamic prediction includes the prediction of the synchronous deviation of each lifting point, the stress distribution of key points of the lifted component, and the trend of hydraulic pressure change in the main lifting system within a future preset time period.

10. The hydraulic synchronous lifting control method driven by a large model according to claim 8, characterized in that, In step S3, the control strategy and parameter optimization includes generating a globally optimal control strategy and dynamically adjusting the PID controller parameters.

Images