A digital twin optimization method and system for open-cut construction over existing subways
Through digital twin model and reinforcement learning, the construction parameters are optimized, and the problem of parameter adjustment dependence in construction of existing subway open-cut contact channels is solved, and a high safety and stability construction process is achieved.
Patent Information
- Application Number
- CN202211499722.7
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2022-11-28
- Publication Date
- 2024-12-06
- Estimated Expiration
- 2042-11-28
AI Technical Summary
In the construction of an open-cut contact channel through existing subways, the adjustment of construction parameters depends on experience and cannot effectively ensure operational safety, which can easily lead to deformation of the subway tunnel structure and affect operational safety.
Through the interaction between the digital twin model and the actual construction scenario, real-time monitoring of data updates and abnormal detection, the construction parameters are optimized using reinforcement learning methods, and the MJS processing method, double-axis mixing pile drilling speed, earth zoning, load capacity and hydraulic breaking parameters are adjusted. Form a closed loop for optimization of construction parameters.
It improves the visualization and automation of construction, provides scientific decision-making basis, enhances construction safety, avoids human decision-making mistakes, and ensures the stability of the subway tunnel structure.
Smart Images

Figure CN116070407B_ABST
Abstract
Description
Technical Field
[0001] The present invention belongs to the field of safety control of open-cut communication channel construction passing through a subway, and relates to a digital twin optimization method and system for open-cut construction passing through an existing subway, and more specifically, to a digital twin optimization method for open-cut construction passing through an existing subway. Background Art
[0002] With the continuous development of my country's urbanization process, cities across the country continue to develop underground space to improve the utilization rate of land construction. Among them, the construction of urban rail transit networks has been steadily advancing, but with it, a large number of projects are approaching and crossing existing operating subways, which not only brings great difficulties to construction, but also easily causes certain impacts on subway tunnels. In severe cases, it may cause excessive deformation of the tunnel structure and affect operational safety. In order to protect the stability of the existing operating subway structure and ensure operational safety, it is necessary to use corresponding technical means to control the construction safely.
[0003] In the open-cut connecting passage project over the existing operating subway, the connecting passage refers to the connecting passage between the basements on both sides of the subway tunnel. Since the connecting passage is closer to the subway tunnel, the subway structure is more affected during construction; it is easy for the subway tunnel structure to sink due to excessive load caused by the stacking of materials and parking of large machinery and equipment above the construction site; similarly, the subway structure may float due to excessive single unloading due to excessive single earth excavation or failure to adjust the earth loading. During the construction process, the adjustment of construction parameters is often based on the work experience of the construction personnel, and the safety of the operation cannot be guaranteed.
[0004] To this end, the present invention proposes a construction parameter optimization method for an open-cut connecting channel passing through an existing operating subway based on digital twins, which can provide a scientific basis for the optimization of construction parameters. Summary of the invention
[0005] In response to the above defects or improvement needs of the prior art, the present invention provides a digital twin optimization method and system for open-cut construction passing through an existing subway, the purpose of which is to optimize and adjust construction parameters through the interaction between the digital twin model and the actual construction scene to ensure the safety of the operation.
[0006] To achieve the above object, according to one aspect of the present invention, a digital twin optimization method for open-cut construction over an existing subway is provided, comprising the following steps:
[0007] S1. Construct a digital twin of the construction of the open-cut connecting channel over the existing operating subway, that is, use the BIM module to construct a digital representation of the physical object in the digital space, so that the construction objects in the real world and the digital twins in the digital space can achieve two-way mapping, data connection and status interaction, and realize data visualization of the construction area including the tunnel;
[0008] S2, obtain monitoring data in real time based on monitoring points on subway tunnels and ground-connected walls, and transmit the monitoring data to the BIM module in real time to achieve real-time update of digital twin data;
[0009] S3. Perform anomaly detection, evaluation and safety classification on the monitoring data of subway tunnels and ground-connected walls, realize safety early warning, and display the anomaly detection, evaluation and safety classification results;
[0010] S4. Optimize the corresponding parameters of the current construction process according to the safety warning situation. Specifically, by inputting the construction plan and construction parameters, using the reinforcement learning method, simulation is performed in the digital twin, and the construction parameters are adjusted and optimized through multiple rounds of reinforcement learning and simulation, and the optimal construction parameters are output;
[0011] S5. Execute the optimal construction parameters on the actual construction object and return to step S2 to form a closed loop of construction parameter optimization based on digital twins.
[0012] Furthermore, in the reinforcement learning method described in step S4, the reinforcement learning environment is set to the BIM model in the digital twin platform, the reinforcement learning state is set to the construction parameter input, and the reinforcement learning action is set to perform a simulation after the construction parameters are changed according to the strategy obtained by reinforcement learning; feedback rewards are given according to the simulation results. If the simulation results show that there is no significant deformation of the tunnel and the ground-connected wall structure, the output is positive feedback; if the deformation is significant, the output is negative feedback.
[0013] Furthermore, in step S4, the adjustment and optimization of the construction parameters includes at least one of the following adjustments: adjustment of the drilling speed of the drill bit in the MJS method and the double-axis mixing pile, adjustment of the earthwork partition size, adjustment of the single excavation volume, adjustment of the loading volume above the steel structure platform, and adjustment of the pressure, flow and stroke of the water pump for breaking the ground-connected wall in water.
[0014] Furthermore, the optimization process of the construction parameters is carried out by adjusting the parameters according to the construction process, and the parameter adjustment methods at different stages are as follows:
[0015] In the stage of strengthening the retaining structure with MJS method and strengthening the foundation with MJS method + biaxial mixing piles:
[0016] According to the construction plan, the bottom of the underground communication channel and some earthwork partitions are reinforced; the damage to the tunnel and the impact of the ground-connected wall are monitored, and visual information is obtained through the safety warning results; the construction process is simulated and reinforced according to the drill bit drilling speed. If the construction parameter simulation results cause too much damage to the tunnel and the ground-connected wall, negative feedback is given to the main body. The main body makes a strategy to reduce the drill bit drilling speed in the MJS method and the double-axis mixing pile and conducts the next round of reinforcement learning until the maximum speed that does not affect the subway tunnel structure is determined, and finally the drill bit drilling speed in the MJS method and the double-axis mixing pile is adjusted;
[0017] During the earthwork excavation phase:
[0018] According to the construction plan, the earthwork excavation area is divided into earthwork partitions of different sizes, and a construction sequence that meets the construction requirements is formulated; tunnel damage and ground-connected wall impacts are monitored according to the construction stage, and visual information is obtained through safety warning results; reinforcement learning and simulation are carried out according to the principle of not affecting the structure and the principle of maximum work efficiency. If the entity receives negative feedback, the entity will make a strategy to reduce the size of the earthwork partition and reduce the single excavation volume and carry out the next round of reinforcement learning until the maximum excavation volume is determined without affecting the tunnel and ground-connected wall structure, and finally the earthwork partition size and single excavation volume are adjusted;
[0019] During the stage of hoisting the prefabricated cushion layer, building the steel structure platform and dynamically adjusting the stacking capacity, as well as the stage of tying and pouring the bottom plate reinforcement, and the stage of measuring the wall and top plate construction:
[0020] Monitor tunnel damage and ground-connected walls according to the construction stage, and obtain visual information through safety warning results; through reinforcement learning, continuously adjust simulation results according to construction parameters, continuously adjust the size of the pile load on the steel structure platform, and output the ups and downs of the subway tunnel as negative feedback. After multiple rounds of reinforcement learning, determine the optimal pile load;
[0021] During the hydraulic demolition of the diaphragm wall:
[0022] Tunnel damage and ground-connected walls are monitored according to the construction stage, and visual information is obtained through safety warning results. Simulation and reinforcement learning are carried out in the simulation model according to pressure, flow and stroke. Under the condition that the pump pressure, flow and stroke are sufficient to break the ground-connected wall, the minimum values of pump pressure, flow and stroke are determined to adjust the hydraulic breaking pressure, flow and stroke to prevent excessive pressure, flow and stroke from causing settlement or horizontal displacement of the ground-connected wall.
[0023] According to another aspect of the present invention, a digital twin optimization system for open-cut construction over an existing subway is provided, comprising a BIM module, a monitoring data input module, a safety warning module, and a construction parameter optimization module;
[0024] The BIM module includes a BIM parametric model, a construction process model and a construction parameter database constructed according to the construction scene of the open-cut communication channel passing through the existing operating subway;
[0025] The BIM parametric model is a parametric model established based on the target building main component data and the construction subway data, wherein the parameters include the structural dimensions of the underground communication channel, the dimensions of the earthwork partition, the structural dimensions of the ground-connected wall, the dimension parameters of the surrounding building complex and the dimension parameters of the tunnel; the monitoring database includes the monitoring data of the tunnel and the ground-connected wall; the construction process model uses the BIM parametric model to simulate the progress of the construction plan of all construction processes; the construction parameter database includes the construction plan data and the daily construction progress data;
[0026] The monitoring data input module is used for collecting and transmitting monitoring data, including the arrangement of tunnel monitoring points and ground-connected wall monitoring points. Specifically, monitoring points are arranged at the top of the tunnel, the waist of the left wall of the tunnel, the waist of the right wall of the tunnel, the left side of the roadbed and the right side of the roadbed, respectively; settlement monitoring points are arranged at the top of the ground-connected wall, and inclination monitoring points are arranged at the bottom of the ground-connected wall; and the monitoring data is transmitted to the BIM module in real time by using IoT sensor equipment to realize real-time update of digital twin platform data;
[0027] The monitoring data include: tunnel settlement, tunnel radial convergence, tunnel horizontal displacement, ground-connected wall settlement and tilt offset;
[0028] The construction parameters include: drilling speed of MJS method and double-shaft mixing pile drilling rig, earthwork block size, single excavation volume, pile load above steel plate, pressure, flow rate and stroke of hydraulically breaking ground-connected wall water pump;
[0029] The safety warning module is used to detect anomalies in the received monitoring data, determine the risk location, perform risk level analysis and assessment, and color-code the risk location according to the risk level for visual display; the safety warning results are fed back to the construction parameter optimization module, and the construction parameter optimization module performs corresponding parameter optimization according to the current construction process;
[0030] The construction parameter optimization module is used to input the construction parameters into the BIM module through reinforcement learning, perform reinforcement learning and simulation, and optimize and adjust the construction parameters according to the reinforcement learning results.
[0031] Furthermore, the process of construction parameter optimization by the construction parameter optimization module is as follows: first, the construction parameters that cause safety problems are input into the BIM module based on the actual construction information and safety warning results; then, the construction parameter optimization module performs reinforcement learning and performs simulation in the BIM module, using the deformation of the subway tunnel and the ground-connected wall structure as feedback, and outputting negative feedback if the deformation is too large. After multiple rounds of reinforcement learning, the construction parameter optimization module obtains the optimal construction parameters to prevent the tunnel and the ground-connected wall structure from deforming; finally, the construction parameter optimization module outputs the updated optimal construction parameters.
[0032] Furthermore, in the reinforcement learning process of the construction parameter optimization module, the reinforcement learning environment is set to the BIM model in the digital twin platform, the reinforcement learning state is set to the construction parameter input, and the reinforcement learning action is set to simulate the changes in the construction parameters according to the strategy obtained by reinforcement learning. Feedback rewards are given according to the simulation results. If the simulation results show that there is no large deformation of the tunnel and the ground-connected wall structure, the output is positive feedback. If the deformation is large, the output is negative feedback.
[0033] Furthermore, the optimization content of the construction parameter optimization module includes at least one of the following adjustments: adjustment of the drilling speed of the drill bit in the MJS method and the double-axis mixing pile, adjustment of the earthwork partition size, adjustment of the single excavation volume, adjustment of the loading volume above the steel structure platform, and adjustment of the pressure, flow and stroke of the water pump for breaking the ground-connected wall in water.
[0034] Furthermore, the construction parameter adjustment methods at different stages are as follows:
[0035] In the stage of strengthening the retaining structure with MJS method and strengthening the foundation with MJS method + biaxial mixing piles:
[0036] According to the construction plan, the bottom of the underground communication channel and some earthwork partitions are reinforced; according to the construction stage, the tunnel damage and the impact of the ground-connected wall are monitored, and visual information is obtained through the safety warning results; through the construction parameter optimization module, the construction under the drill bit drilling speed is simulated and reinforced learning is carried out. If the construction parameter simulation results cause too much damage to the tunnel and the ground-connected wall, negative feedback should be given to the main body. The main body will make a strategy to reduce the drill bit drilling speed in the MJS method and the double-axis mixing pile and carry out the next round of reinforcement learning until the maximum speed that does not affect the subway tunnel structure is determined, and finally the drill bit drilling speed in the MJS method and the double-axis mixing pile is adjusted;
[0037] During the earthwork excavation phase:
[0038] According to the construction plan, the earthwork excavation area is divided into earthwork partitions of different sizes, and a construction sequence that meets the construction requirements is formulated; tunnel damage and ground-connected wall impacts are monitored according to the construction stage, and visual information is obtained through safety warning results; in the construction parameter optimization module, reinforcement learning and simulation are carried out according to the principle of not affecting the structure and the principle of maximum work efficiency. If the entity receives negative feedback, the entity will make a strategy to reduce the size of the earthwork partition and reduce the single excavation volume and carry out the next round of reinforcement learning until the maximum excavation volume is determined without affecting the tunnel and ground-connected wall structure, and finally the earthwork partition size and single excavation volume are adjusted;
[0039] During the stage of hoisting the prefabricated cushion layer, building the steel structure platform and dynamically adjusting the stacking capacity, as well as the stage of tying and pouring the bottom plate reinforcement, and the stage of measuring the wall and top plate construction:
[0040] Monitor tunnel damage and ground-connected walls according to the construction stage, and obtain visual information through safety warning results; in the construction parameter optimization module, through reinforcement learning, the load on the steel structure platform is continuously adjusted according to the construction parameters and simulation results in the simulation model. The floating and sinking of the subway tunnel in the simulation model are output as negative feedback. After multiple rounds of reinforcement learning, the optimal load is determined;
[0041] During the hydraulic demolition of the diaphragm wall:
[0042] Tunnel damage and ground-connected walls are monitored according to the construction stage, and visual information is obtained through safety warning results. The construction parameter optimization module performs simulation and reinforcement learning in the simulation model based on pressure, flow and stroke. While ensuring that the pump pressure, flow and stroke are sufficient to break the ground-connected wall, the minimum values of pump pressure, flow and stroke are determined to adjust the hydraulic breaking pressure, flow and stroke to prevent excessive pressure, flow and stroke from causing settlement or horizontal displacement of the ground-connected wall.
[0043] In general, the above technical solution conceived by the present invention has the following advantages compared with the prior art:
[0044] (1) High degree of visualization: The present invention establishes a digital twin system to remotely display the construction model and progress in real time. In addition, it is equipped with a subway tunnel safety warning module, which can visualize the safety warning results on the digital twin platform and can interact to obtain historical monitoring data.
[0045] (2) High degree of automation: The present invention can adjust the construction parameters based on the existing construction parameters through machine autonomous reinforcement learning and simulation. Compared with adjusting the construction parameters according to manual experience, this method has a higher degree of automation.
[0046] (3) High safety and reliability: Compared with manual experience, this method has a more scientific decision-making basis. The construction parameter optimization involved is more rigorous and reliable than the manual method, which can avoid construction accidents or subway tunnel damage caused by human decision-making errors. BRIEF DESCRIPTION OF THE DRAWINGS
[0047] Figure 1 This is a bird's-eye view of the earthwork blocks above the construction of the open-cut connecting channel that passes through the existing operating subway in this case.
[0048] Figure 2 This is a cross-sectional view of the open-cut connecting passage and subway tunnel that pass through the existing operating subway in this case.
[0049] Figure 3 Architectural diagram of the digital twin optimization method for open-cut construction crossing an existing subway.
[0050] Figure 4 Flowchart of reinforcement learning for construction parameter optimization.
[0051] Figure 5 This is the construction process and construction parameter relationship diagram of the open-cut connecting channel passing through the existing operating subway. DETAILED DESCRIPTION
[0052] In order to make the purpose, technical solutions and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention. In addition, the technical features involved in the various embodiments of the present invention described below can be combined with each other as long as they do not conflict with each other.
[0053] like Figure 1 , Figure 2 As shown, it is a schematic diagram of a construction case of the present invention for passing through an open-cut communication channel of an existing operating subway. The preferred embodiment of the present invention will be introduced below by taking this construction case as an example.
[0054] like Figure 3 As shown, a preferred embodiment of the present invention provides a digital twin optimization method for open-cut construction over an existing subway, and the steps are as follows:
[0055] S1. Build a digital twin platform for the construction of an open-cut connection channel that crosses the existing operating subway, construct a digital representation of physical objects in the digital space, namely the BIM model, to achieve two-way mapping, data connection and status interaction between construction objects in the real world and twins in the digital space, and realize data visualization of the construction area including the tunnel;
[0056] S2. By arranging monitoring points in tunnels and ground-connected walls, the monitoring data is transmitted to the BIM module of the digital twin platform system using IoT sensor equipment, so as to achieve real-time update of the digital twin platform data;
[0057] S3. Based on the monitoring data of subway tunnels and ground-connected walls, the safety warning module of the digital twin platform system performs abnormal detection, evaluation and safety classification on the monitoring data to achieve safety warnings, which are then displayed on the platform.
[0058] S4. The digital twin platform system feeds back the safety warning results to the construction parameter optimization module, and the construction parameter optimization module optimizes the corresponding parameters of the current construction process according to the alarm situation; by inputting the construction plan construction parameters, using the reinforcement learning method, simulation is performed in the digital twin platform simulation model, and according to the output construction parameters, multiple rounds of reinforcement learning and simulation are performed to adjust and optimize, and the optimal construction parameters are output;
[0059] S5. Control and execute the actual construction objects through the optimal construction parameters to meet the actual construction project safety needs and the subway construction requirements, which is conducive to the construction party to dynamically adjust the construction parameters, provide a scientific basis for the adjustment plan, and form an optimization closed loop based on digital twins.
[0060] In order to achieve the above-mentioned objectives, the present invention provides a digital twin system for an open-cut communication channel passing through an existing operating subway, including a BIM module, a monitoring data input module, and a safety warning and construction parameter optimization module.
[0061] The BIM module includes the BIM parametric model of the building construction scenario of the open-cut connecting channel passing through the existing operating subway, the construction process model and the parameter database.
[0062] The BIM parametric model is established based on the main building component data and construction subway data provided by the design institute and the construction party, wherein the parameters include the structural dimensions of the underground communication passage, the dimensions of the earthwork partition, the structural dimensions of the ground-connected wall, the dimension parameters of the surrounding building complex, and the dimension parameters of the tunnel; the monitoring database includes the monitoring data of the tunnel and the ground-connected wall; the construction process model is a progress simulation of the construction plans of all construction processes in the BIM parametric model; the construction parameter database includes the construction plan data, daily construction progress data, etc.
[0063] The monitoring data input module is used for monitoring data collection and transmission. It is necessary to determine the layout of tunnel monitoring points and ground-connected wall monitoring points. Monitoring points are arranged at the top of the tunnel, the waist of the left wall of the tunnel, the waist of the right wall of the tunnel, the left side of the roadbed and the right side of the roadbed. Settlement monitoring points are arranged at the top of the ground-connected wall, and inclination monitoring points are arranged at the bottom of the ground-connected wall. The IoT sensor equipment is used to transmit the monitoring data to the BIM module to realize real-time update of the digital twin platform data.
[0064] The monitoring data include: tunnel settlement (including crown settlement and roadbed settlement), tunnel radial convergence, tunnel horizontal displacement and ground-anchored wall settlement and tilt offset.
[0065] The construction parameters include: drilling speed of MJS method and double-axis mixing pile drilling rig, earthwork block size, single excavation volume, pile load above the steel plate, pressure, flow and stroke of the water pump for hydraulically breaking the ground-connected wall.
[0066] The safety warning and construction parameter optimization modules can be specifically divided into a safety warning module and a construction parameter optimization module; the safety warning module performs anomaly detection on the monitoring data transmitted to the digital twin platform, determines the risk location, conducts risk level analysis and assessment, and color-codes the risk locations according to the risk level, which are visualized on the digital twin platform; the platform feeds back the safety warning results to the construction parameter optimization module, and the construction parameter optimization module performs corresponding parameter optimization based on the current construction process.
[0067] The construction parameter optimization module uses reinforcement learning to input the construction parameters into the digital twin platform simulation model, conducts reinforcement learning and simulation, and optimizes and adjusts the construction parameters according to the reinforcement learning results. Specifically, the construction parameter optimization process first automatically inputs the construction parameters with safety problems into the digital twin platform simulation model based on the actual construction information and safety warning results; then, the construction parameter optimization module conducts reinforcement learning and simulation in the simulation model, using the deformation of the subway tunnel and ground-connected wall structure as feedback, and outputs negative feedback if the deformation is too large. After multiple rounds of reinforcement learning, the module can obtain the optimal construction parameters to prevent the deformation of the tunnel and ground-connected wall structure; finally, the module outputs the updated optimal construction parameters.
[0068] During the reinforcement learning process, the environment represents the BIM model in the digital twin platform. The reinforcement learning state is set to the construction parameter input. The reinforcement learning action is set to the simulation after the construction parameter change according to the reinforcement learning strategy. Feedback rewards are given according to the simulation results. If the simulation results show that the tunnel and the ground-connected wall structure have no large deformation, the output is positive feedback. If the deformation is large, the output is negative feedback. The process is as follows: Figure 4 As shown, there are specifically the following steps:
[0069] (S1) inputting construction parameters into the construction environment model;
[0070] (S2) resetting the construction environment model and returning to the first state;
[0071] (S3) implementing an entity action according to the strategy, simulating the construction environment model, and causing a change in the construction environment model;
[0072] (S4) The environment gives feedback and rewards to the entity to determine whether the tunnel and the ground-connected wall have undergone significant deformation. If so, negative feedback is given to the entity and the process returns to step (S2). If not, positive feedback is given.
[0073] The construction parameters are optimized as follows Figure 5 Shown are the adjustment of the drilling speed of the drill bit in the MJS method and double-axis mixing piles, the earthwork partition size, the adjustment of the single excavation volume, the adjustment of the load above the steel structure platform, and the pressure, flow and stroke of the water pump for breaking the ground-connected wall in water.
[0074] The construction parameter optimization process adjusts the parameters according to the construction process. The construction parameter adjustment methods at different construction stages are as follows:
[0075] In the stage of strengthening the retaining structure with MJS method and strengthening the foundation with MJS method + biaxial mixing piles:
[0076] like Figure 1 , Figure 2 As shown in the figure, according to the construction plan, the bottom of the underground communication channel and the earthwork partitions of ①-c, ①-e, ②-c, ②-e, ③-c, ③-e, ④-a, ④-b, ④-c, ④-d, ④-e, ⑤-a, ⑤-b, ⑤-c, ⑤-d, and ⑤-e need to be reinforced; the damage to the tunnel and the impact of the ground-connected wall are monitored according to the construction stage, and visual information is obtained in the digital twin platform through the safety warning results; according to the construction parameter optimization module, the construction under the drill bit drilling speed is simulated, and reinforcement learning is performed in the simulation model of the digital twin platform. If the construction parameter simulation results cause too much damage to the tunnel and the ground-connected wall, negative feedback should be given to the main body, and the main body will make a strategy to reduce the drill bit drilling speed in the MJS method and the double-axis mixing pile and carry out the next round of reinforcement learning until the maximum speed that does not affect the subway tunnel structure is determined, and finally the adjustment of the drill bit drilling speed in the MJS method and the double-axis mixing pile is realized.
[0077] During the earthwork excavation phase:
[0078] like Figure 1 , Figure 2 As shown, according to the construction plan, the earthwork excavation area is divided into 25 (5*5) earthwork partitions, and the construction sequence is as follows:
[0079] Phase 1: ①-a→②-a→③-a→④-a→⑤-a
[0080] The second stage: ①-b→②-b→③-b→④-b→⑤-b
[0081] The third stage: ①-c→②-c→③-c→④-c→⑤-c
[0082] The fourth stage: ①-d→②-d→③-d→④-d→⑤-d
[0083] The fifth stage: ①-e→②-e→③-e→④-e→⑤-e
[0084] Tunnel damage and ground-anchored wall impacts are monitored according to the construction phase, and visual information is obtained in the digital twin platform through safety warning results. In the construction parameter optimization module, reinforcement learning and simulation are carried out according to the principle of not affecting the structure and the principle of maximizing work efficiency. If the entity receives negative feedback, it will make a strategy to reduce the earthwork partition size and the single excavation volume and carry out the next round of reinforcement learning until the maximum excavation volume is determined without affecting the tunnel and ground-anchored wall structure, and finally the earthwork partition size and single excavation volume are adjusted.
[0085] During the stage of hoisting the prefabricated cushion layer, building the steel structure platform and dynamically adjusting the stacking capacity, as well as the stage of tying and pouring the bottom plate reinforcement, and the stage of measuring the wall and top plate construction:
[0086] Tunnel damage and ground-anchored walls are monitored according to the construction stage, and visual information is obtained in the digital twin platform through safety warning results. In the construction parameter optimization module, reinforcement learning is used to continuously adjust the load on the steel structure platform in the simulation model according to the construction parameters and simulation results. The floating and sinking of the subway tunnel in the simulation model are output as negative feedback. After multiple rounds of reinforcement learning, the optimal load is determined.
[0087] During the hydraulic demolition of the diaphragm wall:
[0088] Tunnel damage and ground-connected walls are monitored according to the construction stage, and visual information is obtained in the digital twin platform through safety warning results. In the construction parameter optimization module, simulation and reinforcement learning are carried out in the simulation model according to pressure, flow and stroke. While ensuring that the pump pressure, flow and stroke are sufficient to break the ground-connected wall, their minimum values are determined to adjust the hydraulic breaking pressure, flow and stroke to prevent excessive pressure, flow and stroke from causing settlement or horizontal displacement of the ground-connected wall.
[0089] It will be easily understood by those skilled in the art that the above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions and improvements made within the spirit and principles of the present invention should be included in the protection scope of the present invention.
Claims
1. A digital twin optimization method for open-cut construction over an existing subway, characterized in that: The steps include: S1. Construct a digital twin of the construction of the open-cut connecting channel over the existing operating subway, that is, use the BIM module to construct a digital representation of the physical object in the digital space, so that the construction objects in the real world and the digital twins in the digital space can achieve two-way mapping, data connection and status interaction, and realize data visualization of the construction area including the tunnel; S2, obtain monitoring data in real time based on monitoring points on subway tunnels and ground-connected walls, and transmit the monitoring data to the BIM module in real time to achieve real-time update of digital twin data; S3. Perform anomaly detection, evaluation and safety classification on the monitoring data of subway tunnels and ground-connected walls, realize safety early warning, and display the anomaly detection, evaluation and safety classification results; S4. Optimize the corresponding parameters of the current construction process according to the safety warning situation. Specifically, by inputting the construction plan and construction parameters, using the reinforcement learning method, simulation is performed in the digital twin, and the construction parameters are adjusted and optimized through multiple rounds of reinforcement learning and simulation, and the optimal construction parameters are output; S5. Execute the optimal construction parameters on the actual construction object and return to step S2 to form a closed loop of construction parameter optimization based on digital twins.
2. The digital twin optimization method for open-cut construction over an existing subway as claimed in claim 1, characterized in that: In the reinforcement learning method described in step S4, the reinforcement learning environment is set to the BIM model in the digital twin platform, the reinforcement learning state is set to the construction parameter input, and the reinforcement learning action is set to perform a simulation after the construction parameters are changed according to the strategy obtained by reinforcement learning; feedback rewards are given according to the simulation results. If the simulation results show that there is no large deformation of the tunnel and the ground-connected wall structure, the output is positive feedback; if the deformation is large, the output is negative feedback.
3. The digital twin optimization method for open-cut construction over an existing subway as claimed in claim 1 is characterized in that: In step S4, the adjustment and optimization of the construction parameters include at least one of the following adjustments: adjustment of the drilling speed of the drill bit in the MJS method and the double-axis mixing pile, adjustment of the earthwork partition size, adjustment of the single excavation volume, adjustment of the loading volume above the steel structure platform, and adjustment of the pressure, flow and stroke of the water pump for breaking the ground-connected wall in water.
4. The digital twin optimization method for open-cut construction over an existing subway as claimed in claim 3 is characterized in that: The optimization process of the construction parameters is carried out according to the construction process. The parameter adjustment methods at different stages are as follows: In the stage of strengthening the retaining structure with MJS method and strengthening the foundation with MJS method + biaxial mixing piles: According to the construction plan, the bottom of the underground communication channel and some earthwork partitions are reinforced; the damage to the tunnel and the impact of the ground-connected wall are monitored, and visual information is obtained through the safety warning results; the construction process is simulated and reinforced according to the drill bit drilling speed. If the construction parameter simulation results cause too much damage to the tunnel and the ground-connected wall, negative feedback is given to the main body. The main body makes a strategy to reduce the drill bit drilling speed in the MJS method and the double-axis mixing pile and conducts the next round of reinforcement learning until the maximum speed that does not affect the subway tunnel structure is determined, and finally the drill bit drilling speed in the MJS method and the double-axis mixing pile is adjusted; During the earthwork excavation phase: According to the construction plan, the earthwork excavation area is divided into earthwork partitions of different sizes, and a construction sequence that meets the construction requirements is formulated; tunnel damage and ground-connected wall impacts are monitored according to the construction stage, and visual information is obtained through safety warning results; reinforcement learning and simulation are carried out according to the principle of not affecting the structure and the principle of maximum work efficiency. If the entity receives negative feedback, the entity will make a strategy to reduce the size of the earthwork partition and reduce the single excavation volume and carry out the next round of reinforcement learning until the maximum excavation volume is determined without affecting the tunnel and ground-connected wall structure, and finally the earthwork partition size and single excavation volume are adjusted; During the stage of hoisting the prefabricated cushion layer, building the steel structure platform and dynamically adjusting the stacking capacity, as well as the stage of tying and pouring the bottom plate reinforcement, and the stage of measuring the wall and top plate construction: Monitor tunnel damage and ground-connected walls according to the construction stage, and obtain visual information through safety warning results; through reinforcement learning, continuously adjust simulation results according to construction parameters, continuously adjust the size of the pile load on the steel structure platform, and output the ups and downs of the subway tunnel as negative feedback. After multiple rounds of reinforcement learning, determine the optimal pile load; During the hydraulic demolition of the diaphragm wall: Tunnel damage and ground-connected walls are monitored according to the construction stage, and visual information is obtained through safety warning results. Simulation and reinforcement learning are carried out in the simulation model according to pressure, flow and stroke. Under the condition that the pump pressure, flow and stroke are sufficient to break the ground-connected wall, the minimum values of pump pressure, flow and stroke are determined to adjust the hydraulic breaking pressure, flow and stroke to prevent excessive pressure, flow and stroke from causing settlement or horizontal displacement of the ground-connected wall.
5. A digital twin optimization system for open-cut construction over an existing subway, characterized in that: Including BIM module, monitoring data input module, safety warning module and construction parameter optimization module; The BIM module includes a BIM parametric model, a construction process model and a construction parameter database constructed according to the construction scene of the open-cut communication channel passing through the existing operating subway; The BIM parametric model is a parametric model established based on the target building main component data and the construction subway data, wherein the parameters include the structural dimensions of the underground communication channel, the dimensions of the earthwork partition, the structural dimensions of the ground-connected wall, the dimension parameters of the surrounding building complex and the dimension parameters of the tunnel; the monitoring database includes the monitoring data of the tunnel and the ground-connected wall; the construction process model uses the BIM parametric model to simulate the progress of the construction plan of all construction processes; the construction parameter database includes the construction plan data and the daily construction progress data; The monitoring data input module is used for collecting and transmitting monitoring data, including the arrangement of tunnel monitoring points and ground-connected wall monitoring points. Specifically, monitoring points are arranged at the top of the tunnel, the waist of the left wall of the tunnel, the waist of the right wall of the tunnel, the left side of the roadbed and the right side of the roadbed, respectively; settlement monitoring points are arranged at the top of the ground-connected wall, and inclination monitoring points are arranged at the bottom of the ground-connected wall; and the monitoring data is transmitted to the BIM module in real time by using IoT sensor equipment to realize real-time update of digital twin platform data; The monitoring data include: tunnel settlement, tunnel radial convergence, tunnel horizontal displacement, ground-connected wall settlement and tilt offset; The construction parameters include: drilling speed of MJS method and double-shaft mixing pile drilling rig, earthwork block size, single excavation volume, pile load above steel plate, pressure, flow rate and stroke of hydraulically breaking ground-connected wall water pump; The safety warning module is used to detect anomalies in the received monitoring data, determine the risk location, perform risk level analysis and assessment, and color-code the risk location according to the risk level for visual display; the safety warning results are fed back to the construction parameter optimization module, and the construction parameter optimization module performs corresponding parameter optimization according to the current construction process; The construction parameter optimization module is used to input the construction parameters into the BIM module through reinforcement learning, perform reinforcement learning and simulation, and optimize and adjust the construction parameters according to the reinforcement learning results.
6. A digital twin optimization system for open-cut construction over an existing subway as claimed in claim 5, characterized in that: The process of construction parameter optimization by the construction parameter optimization module is as follows: first, the construction parameters that cause safety problems are input into the BIM module according to the actual construction information and safety warning results; then, the construction parameter optimization module performs reinforcement learning and performs simulation in the BIM module, taking the deformation of the subway tunnel and the ground-connected wall structure as feedback, and outputting negative feedback if the deformation is too large. After multiple rounds of reinforcement learning, the construction parameter optimization module obtains the optimal construction parameters to prevent the deformation of the tunnel and the ground-connected wall structure; finally, the construction parameter optimization module outputs the updated optimal construction parameters.
7. The digital twin optimization system for open-cut construction over an existing subway as claimed in claim 5, characterized in that: During the reinforcement learning process of the construction parameter optimization module, the construction environment is represented as a BIM model in the digital twin platform, the reinforcement learning state is set to the construction parameter input, and the reinforcement learning action is set to simulate the changes in the construction parameters according to the strategy obtained by reinforcement learning. Feedback rewards are given according to the simulation results. If the simulation results show that there is no large deformation of the tunnel and ground-connected wall structure, the output is positive feedback. If the deformation is large, the output is negative feedback.
8. The digital twin optimization system for open-cut construction over an existing subway as claimed in claim 5, characterized in that: The optimization content of the construction parameter optimization module includes at least one of the following adjustments: adjustment of the drilling speed of the drill bit in the MJS method and double-axis mixing piles, adjustment of the earthwork partition size, adjustment of the single excavation volume, adjustment of the load above the steel structure platform, and adjustment of the pressure, flow and stroke of the water pump for breaking the ground-connected wall in water.
9. A digital twin optimization system for open-cut construction over an existing subway as claimed in claim 8, characterized in that: The construction parameter adjustment methods at different stages are as follows: In the stage of strengthening the retaining structure with MJS method and strengthening the foundation with MJS method + biaxial mixing piles: According to the construction plan, the bottom of the underground communication channel and some earthwork partitions are reinforced; according to the construction stage, the tunnel damage and the impact of the ground-connected wall are monitored, and visual information is obtained through the safety warning results; through the construction parameter optimization module, the construction under the drill bit drilling speed is simulated and reinforced learning is carried out. If the construction parameter simulation results cause too much damage to the tunnel and the ground-connected wall, negative feedback should be given to the main body. The main body will make a strategy to reduce the drill bit drilling speed in the MJS method and the double-axis mixing pile and carry out the next round of reinforcement learning until the maximum speed that does not affect the subway tunnel structure is determined, and finally the drill bit drilling speed in the MJS method and the double-axis mixing pile is adjusted; During the earthwork excavation phase: According to the construction plan, the earthwork excavation area is divided into earthwork partitions of different sizes, and a construction sequence that meets the construction requirements is formulated; tunnel damage and ground-connected wall impacts are monitored according to the construction stage, and visual information is obtained through safety warning results; in the construction parameter optimization module, reinforcement learning and simulation are carried out according to the principle of not affecting the structure and the principle of maximum work efficiency. If the entity receives negative feedback, the entity will make a strategy to reduce the size of the earthwork partition and reduce the single excavation volume and carry out the next round of reinforcement learning until the maximum excavation volume is determined without affecting the tunnel and ground-connected wall structure, and finally the earthwork partition size and single excavation volume are adjusted; During the stage of hoisting the prefabricated cushion layer, building the steel structure platform and dynamically adjusting the stacking capacity, as well as the stage of tying and pouring the bottom plate reinforcement, and the stage of measuring the wall and top plate construction: Monitor tunnel damage and ground-connected walls according to the construction stage, and obtain visual information through safety warning results; in the construction parameter optimization module, through reinforcement learning, the load on the steel structure platform is continuously adjusted according to the construction parameters and simulation results in the simulation model. The floating and sinking of the subway tunnel in the simulation model are output as negative feedback. After multiple rounds of reinforcement learning, the optimal load is determined; During the hydraulic demolition of the diaphragm wall: Tunnel damage and ground-connected walls are monitored according to the construction stage, and visual information is obtained through safety warning results. The construction parameter optimization module performs simulation and reinforcement learning in the simulation model based on pressure, flow and stroke. While ensuring that the pump pressure, flow and stroke are sufficient to break the ground-connected wall, the minimum values of pump pressure, flow and stroke are determined to adjust the hydraulic breaking pressure, flow and stroke to prevent excessive pressure, flow and stroke from causing settlement or horizontal displacement of the ground-connected wall.
Citation Information
Patent Citations
BIM modularization modeling method and BIM modularization modeling system
CN107239631A
Quick and refined mountain tunnel BIM building method
CN108563851A
Information model expression method for subway section construction method
CN108920820A
Method and system for realizing tunnel digital display based on BIM and 3DGIS technology system
CN112347548A
Proximity engineering construction load test method and system based on digital twinning
CN114139270A
