BIM intelligent assembly system applied to node construction

Abstract

The invention discloses a BIM intelligent assembly system applied to node construction, and belongs to the technical field of node construction. The BIM intelligent assembly system comprises the following modules: a node-level BIM parameterized model library; a factory-end intelligent manufacturing module; the field end intelligent assembly module comprises a reusable node sensing layer composed of UWB, IMU and a visual fusion sensing suite; an edge decision-making layer in which a lightweight Transform model is embedded; the shape memory alloy lock pin and the self-locking and self-adjusting execution layer of the piezoelectric micro-displacement sliding table are integrated on the node connecting plate; and a block chain quality file sub-module. The parameterized node group is directly issued to the steel bar hoop bending robot, the sleeve welding robot and the 3D printing mold, the space coordinates of steel bars, sleeves and embedded parts can be locked within the millimeter-level range in the factory stage, and the six-degree-of-freedom postures of the nodes are captured in real time during on-site hoisting; once deviation is found, the edge AI gives out a posture adjusting instruction immediately, and correction is completed through an SMA lock pin and a piezoelectric micro-displacement sliding table.

Description

TECHNICAL FIELD

[0001] The application belongs to the technical field of node construction, and particularly relates to a BIM intelligent assembly system applied to node construction. BACKGROUND

[0002] In high-rise fabricated residential buildings and large public buildings, the nodes where beams, columns, shear walls and floors intersect are not only the key path for transmitting internal forces, but also the most densely intersected parts of multiple specialties such as mechanical and electrical pipelines, curtain wall embedded parts, fireproof plugging and the like. The traditional method usually provides two-dimensional node details by a design institute, precast embedded parts in a factory, and then workers complete the steel bar hole matching, sleeve positioning and grouting by relying on a tape measure, a level and a crowbar. Since the two-dimensional drawing cannot completely express the three-dimensional spatial relationship, the collision conflict is often discovered when the component is transported under the tower crane. The workers judge the steel bar gap by eyesight, and it is easy to cause common quality problems such as main reinforcement deviation, sleeve grouting leakage and grouting incompleteness. Once the error accumulates, the site can only adopt extensive measures such as flame cutting, steel plate repair welding or secondary pouring of concrete to remedy, which not only delays the construction period but also produces additional carbon emissions.

[0003] The existing technical system still has significant defects: on the one hand, the information fault between the two-dimensional drawing and the three-dimensional entity has not been fundamentally solved. The traditional two-dimensional node detail drawing can express single professional information, but cannot dynamically associate the spatial requirements of multiple specialties such as mechanical and electrical pipelines and curtain wall embedded parts, resulting in the frequent occurrence of the "space fighting" phenomenon when the prefabricated component is hoisted on site. On the other hand, the existing BIM technology application still stays at the "visualization display" level. There is a lack of standardized interface protocol between the model data and the intelligent manufacturing equipment, resulting in the common existence of the "model-entity" two-skin phenomenon. In addition, the existing quality control system has the dual problems of "data island" and "process black box". The quality detection data is scattered in the paper records, Excel tables and independent software systems of different participants, and lacks a unified data governance framework. Therefore, a BIM intelligent assembly system applied to node construction is proposed. SUMMARY

[0004] The purpose of the application is to provide a BIM intelligent assembly system applied to node construction to solve the problems in the background art.

[0005] To achieve the above purpose, the application provides the following technical scheme: comprising the following modules:

[0006] Node-level BIM parameterized model library: composed of LOD400 node families, and embedded with collision rules, steel bar avoidance algorithms and tolerance limits in the families;

[0007] Factory-side intelligent manufacturing module: data intercommunication with the node-level BIM parameterized model library, driving the steel bar hoop robot, sleeve automatic welding machine and 3D printing positioning mold to manufacture node embedded parts with an accuracy of ≤0.5 mm.

[0008] On-site end intelligent assembly module: including a node perception layer composed of a reusable UWB+IMU+vision fusion sensor suite, which obtains the six-degree-of-freedom pose of the node steel bar and sleeve in real time; an edge decision layer embedded with a lightweight Transformer model, which predicts the node assemblability based on the real-time pose and BIM theoretical coordinates and outputs the pose adjustment strategy; a self-locking and self-adjusting execution layer integrated with a shape memory alloy lock pin and a piezoelectric micro-displacement slide table on the node connecting plate, which automatically completes the steel bar hole matching, sleeve fine adjustment and temporary locking according to the pose adjustment strategy;

[0009] Blockchain quality archive sub-module: the whole process data from node blanking in the factory to grouting completion are written into the alliance chain in the form of BIM-hash, realizing node-level quality lifelong traceability;

[0010] As a further preferred embodiment of the present technical solution: the node-level BIM parameterized model includes a node type automatic classifier for automatically matching the optimal node family according to the structure calculation results and a tolerance allowance dynamic adjuster for real-time correcting the manufacturing tolerance according to the on-site environment temperature and humidity and component age;

[0011] As a further preferred embodiment of the present technical solution: the real-time visual closed-loop feedback unit of the steel bar hoop bending robot end effector in the factory end intelligent manufacturing module has an actual bending error ≤0.3mm, and the laser weld tracking unit of the sleeve automatic welding machine in the factory end intelligent manufacturing module ensures that the sleeve coaxiality is ≤0.2mm;

[0012] As a further preferred embodiment of the present technical solution: the UWB+IMU+vision fusion sensor suite in the node perception layer adopts a magnetic quick-release structure; the Transformer model of the edge decision layer is quantized and pruned, with a model size ≤10MB, an inference delay ≤100ms, and running in an edge computing box in the crane cockpit; the driving current of the shape memory alloy lock pin of the self-locking and self-adjusting execution layer is 0.8-1.2A, the locking force is ≥3kN, and the response time is ≤200ms; the stroke of the piezoelectric micro-displacement slide table is ±2mm, the closed-loop resolution is ≤0.5μm, and the maximum load is ≥500N;

[0013] As a further preferred embodiment of the present technical solution: the blockchain quality archive sub-module adopts BIM-hash, which is: combining the GUID of the node BIM model object, the manufacturing batch number, the Hash value of the on-site sensing data and the timestamp to generate a Merkle root and write it into the alliance chain;

[0014] As a further preferred of the technical solution: further comprising a digital twin synchronization engine, after the node assembly is completed, automatically write back the field measured coordinates to the cloud BIM model, update the deviation of model coordinates and entity coordinates ≤0.1mm;

[0015] As a further preferred of the technical solution: comprising the following steps:

[0016] S1, calling node-level parameterized family in BIM platform, completing collision checking and reinforcement avoidance optimization, and generating manufacturing instructions;

[0017] S2, the factory end completes the intelligent manufacturing of the node embedded part according to the manufacturing instructions, and implants an RFID identity tag in the embedded part;

[0018] S3, before hoisting, automatically match the BIM node instance through RFID identification, and send it to the edge decision layer;

[0019] S4, in the hoisting process, the node perception layer collects the reinforcement / sleeve pose in real time, and the edge decision layer outputs the pose adjustment strategy;

[0020] S5, the self-locking and self-adjusting execution layer completes the reinforcement hole matching, sleeve fine adjustment and temporary locking according to the strategy;

[0021] S6, after grouting is completed, the blockchain quality file submodule chains the whole process data, and the digital twin synchronization engine updates the BIM model;

[0022] As a further preferred of the technical solution: in the step S4, when the predicted assembly score is >95%, the crane micro-motion mode is automatically triggered, and millimeter-level translation or rotation instructions are sent to the crane PLC through the edge computing box;

[0023] As a further preferred of the technical solution: in the step S6, when the node grouting strength real-time monitoring value is lower than the design value by 90%, the next process is automatically suspended through the blockchain smart contract, and a warning is pushed to the supervision APP;

[0024] As a further preferred of the technical solution: the field data is uploaded to the cloud through a 5G network, and the end-to-end delay is ≤20ms.

[0025] Compared with the prior art, the beneficial effects of the present application are:

[0026] 1、The application can lock the spatial coordinates of the steel bars, sleeves and embedded parts in the millimeter level range in the factory stage, capture the six-degree-of-freedom attitude of the nodes in real time during the hoisting on site, and compare with the cloud model in seconds, so that the edge AI can give the attitude adjustment instruction immediately once the deviation is found, and the correction can be completed by the SMA lock pin and the piezoelectric micro-displacement slide table, so that the whole process does not need repeated measurement and knocking by manual operation, and the consistency with the design intention is ensured.

[0027] 2、The application can realize the "positioning and locking" by hoisting only once through the data through of "model-factory-site", so that a large amount of hole matching and fine adjustment work can be completed in the ground factory, the single node operation time is significantly shortened, the edge decision module can monitor the safety indexes such as wind speed, load and node stress in real time, and the crane can be automatically triggered to slow down, pause or retreat once the abnormality is detected, so that the risky operation under the overload or strong wind condition is avoided, and the safety is improved. BRIEF DESCRIPTION OF DRAWINGS

[0028] Figure 1 It is a structural diagram of a BIM intelligent assembly system applied to node construction;

[0029] Figure 2 It is a flowchart of a BIM intelligent assembly system applied to node construction. DETAILED DESCRIPTION

[0030] The technical solutions in the embodiments of the application will be clearly and completely described below with reference to the drawings in the embodiments of the application. Obviously, the described embodiments are only part of the embodiments of the application, rather than all the embodiments of the application. Based on the embodiments in the application, all other embodiments obtained by those skilled in the art without creative labor are within the protection scope of the application.

[0031] EMBODIMENT

[0032] Please refer to Figure 1 - Figure 2 The application provides a technical solution comprising the following modules:

[0033] The node-level BIM parameterized model library is composed of LOD400 node families, and the collision rules, steel bar avoidance algorithms and tolerance limits are embedded in the families, so that the "design is manufacturing" is realized through the built-in LOD400 node families and collision rules, and the information distortion caused by the later two-dimensional change is avoided;

[0034] The factory-end intelligent manufacturing module is in data intercommunication with the node-level BIM parameterized model library, drives the steel bar bending robot, sleeve automatic welding machine and 3D printing positioning mold, and manufactures the node embedded parts with the accuracy of ≤0.5 mm.

[0035] On-site end intelligent assembly module: including a node perception layer composed of a reusable UWB+IMU+vision fusion sensor suite, which can obtain the six-degree-of-freedom pose of the node steel bar and sleeve in real time; an edge decision layer embedded with a lightweight Transformer model, which can predict the node assemblability based on the real-time pose and BIM theoretical coordinates and output the pose adjustment strategy; a self-locking and self-adjusting execution layer integrated with a shape memory alloy lock pin and a piezoelectric micro-displacement slide on the node connecting plate, which can automatically complete the steel bar hole alignment, sleeve fine adjustment and temporary locking according to the pose adjustment strategy;

[0036] Blockchain quality archive sub-module: the whole process data from factory blanking to grouting completion of the node is written into the alliance chain in the form of BIM-hash, realizing the node-level quality lifelong traceability;

[0037] In this embodiment, specifically: the node-level BIM parameterized model includes a node type automatic classifier for automatically matching the optimal node family according to the structure calculation results and a tolerance allowance dynamic adjuster for real-time correction of manufacturing tolerances according to the on-site environment temperature and humidity and component age, and the node type automatic classifier and the tolerance dynamic adjuster work cooperatively, so that the same set of family library can be adapted to different seismic fortification intensity areas without repeated modeling;

[0038] In this embodiment, specifically: the real-time visual closed-loop feedback unit of the steel bar hoop bending robot end effector in the factory end intelligent manufacturing module can ensure that the actual bending error is ≤0.3mm, and the laser weld seam tracking unit of the sleeve automatic welding machine in the factory end intelligent manufacturing module can ensure that the sleeve coaxiality is ≤0.2mm, and the visual closed-loop feedback and laser weld seam tracking units form a "double closed loop", which can ensure zero defects of bending and welding even if there is an oxide skin on the surface of the steel bar;

[0039] In this embodiment, specifically: the UWB+IMU+vision fusion sensing suite in the node perception layer adopts a magnetic quick-release structure; the Transformer model of the edge decision layer is pruned and quantized, the model size is ≤10MB, the inference delay is ≤100ms, and it runs in the edge computing box in the crane cabin; the driving current of the shape memory alloy lock pin of the self-locking self-adjusting execution layer is 0.8-1.2A, the locking force is ≥3kN, the response time is ≤200ms, the stroke of the piezoelectric micro-displacement slide table is ±2mm, the closed-loop resolution is ≤0.5μm, and the maximum load is ≥500N; the magnetic quick-release structure allows the sensor to be used with the spreader, reducing the one-time investment cost and improving the site turnover efficiency; the lightweight Transformer model continuously learns incrementally based on the field data set, and the node recognition accuracy continuously improves as the project progresses; the lock pin triggers phase change using the ambient temperature, without the need for additional heating devices, reducing the complexity of cable layout; the piezoelectric slide table is equipped with a closed-loop grating ruler, which can be used for secondary review before grouting, ensuring that the reinforcement cover thickness meets the specification requirements;

[0040] In this embodiment, specifically: the blockchain quality archive submodule adopts BIM-hash, which is: combining the GUID of the node BIM model object, the manufacturing batch number, the Hash value of the field sensing data, and the timestamp to generate a Merkle root and write it to the consortium chain. BIM-hash binds the model GUID with the sensing data Hash value, ensuring that any subsequent tampering will be immediately detected by the on-chain nodes;

[0041] In this embodiment, specifically: it also includes a digital twin synchronization engine that automatically writes the field measured coordinates back to the cloud BIM model after node assembly, updates the model coordinate and entity coordinate deviation ≤0.1mm, and supports IFC4.3 format, which can directly interface with the city-level CIM platform to realize regional prefabricated quality supervision;

[0042] In this embodiment, specifically: the following steps are included:

[0043] S1, call the node-level parameterized family in the BIM platform, complete the collision check and reinforcement avoidance optimization, and generate the manufacturing instructions;

[0044] S2, the factory side completes the intelligent manufacturing of the node embedded part according to the manufacturing instructions, and implants an RFID identity tag in the embedded part;

[0045] S3, before hoisting, automatically match the BIM node instance through RFID identification and issue it to the edge decision layer; the RFID identity tag automatically triggers the "zero-touch" verification during the hoisting process to prevent different batches of components from being mixed;

[0046] S4, during hoisting, the node perception layer collects the reinforcement / sleeve pose in real time, and the edge decision layer outputs the pose adjustment strategy;

[0047] S5, the self-locking self-adjusting execution layer completes the hole matching of the steel bars, sleeve fine adjustment and temporary locking according to the strategy;

[0048] S6, after grouting is completed, the blockchain quality file sub-module chains the whole process data, and a digital twin synchronous engine updates a BIM model;

[0049] In the embodiment, specifically: in step S4, when the predicted assemblability score is greater than 95%, the crane micro-motion mode is automatically triggered, millimeter-level translation or rotation instructions are sent to the crane PLC through the edge computing box, and the crane micro-motion mode is controlled through the PLC to realize millimeter-level translation and rotation of the component in the air, thereby reducing manual prying;

[0050] In the embodiment, specifically: in step S6, when the real-time monitoring value of the node grouting strength is lower than 90% of the design value, the next process is automatically suspended through the blockchain smart contract, and a warning is pushed to the supervision APP, and the suspendable mechanism of the blockchain smart contract provides the supervision unit with the "one-key stop" permission;

[0051] In the embodiment, specifically: the on-site data is uploaded to the cloud through the 5G network, and the end-to-end delay is less than or equal to 20 ms.

[0052] Working principle or structural principle: In the design stage, the structural engineer calls the "beam-column edge node parameterized family" on the BIM platform, inputs the key parameters such as beam height, column section, steel grade, etc., the system automatically runs the collision check and steel avoidance algorithm, generates the LOD400 model and locks all coordinates; the factory uses the LOD400 model to directly connect the steel bending hoop robot through API, completes the main reinforcement, hoop reinforcement cutting and bending, the sleeve automatic welding machine welds the sleeve according to the model coordinates, the 3D printing positioning mold is printed synchronously, ensures that the space error of the embedded part is within the allowable range, the system takes the BIM model as the core, generates the LOD400 level node family library through the parameterized design method, embeds the collision rules, steel avoidance algorithm and tolerance tolerance dynamic adjustment mechanism, the designer inputs the structural parameters through the interactive interface, the system automatically generates the node model meeting the specification requirements, and outputs the manufacturing instructions, this process realizes the standardization and automation of design data, avoids the errors and omissions caused by manual drawing, after receiving the manufacturing instructions, the factory end identifies the embedded part through the RFID identity tag system, and establishes the mapping relationship with the virtual components in the BIM model.The high-precision manufacturing equipment completes key processes such as steel bar bending, sleeve welding and embedded part positioning according to geometric data and process parameters in the BIM model, and real-time visual closed-loop feedback units and laser weld tracking units ensure the precision and quality of the manufacturing process, realizing the manufacturing goal of "design as obtained". Before the component is loaded onto the vehicle, the RFID identity tag is bound to the model GUID, and the temperature, humidity and vibration overrun data during transportation are automatically returned to the cloud. During the on-site hoisting process, the tower crane hovers the precast beam 50 cm above the node, and the on-site sensing kit captures the relative pose of the beam end steel bar and the column end sleeve in real time. The edge AI judges the assemblability within 100 ms and outputs Δx, Δy and Δθz, which are extended through the SMA lock pin, and the piezoelectric sliding table completes the fine adjustment, and the steel bar is successfully one-time aligned, and the worker only needs to do grouting and temporary support. The UWB+IMU+vision fusion sensing kit collects node pose data in real time, and a lightweight Transformer model is run through the edge computing box for real-time processing. The model captures the spatiotemporal correlation of the pose data through the self-attention mechanism, generates a high-precision pose adjustment strategy, and the self-locking and self-adjusting execution layer drives the piezoelectric micro-displacement sliding table and the shape memory alloy lock pin to complete the steel bar alignment, sleeve fine adjustment and temporary locking, etc. The whole pose adjustment process does not require manual intervention, realizing the automation and intelligentization of assembly. Finally, the grouting fullness sensor uploads data in real time, the system generates a "BIM-hash" block and writes it into the consortium chain, and through the digital twin synchronization engine, the on-site measured coordinates are written back to the cloud model. The system builds a quality file submodule through blockchain technology, stores the whole process data of the node from design to construction on the consortium chain in an unalterable form, the BIM-hash mechanism ensures the uniqueness and authenticity of the data, the intelligent contract function realizes the automatic execution of quality control, the digital twin synchronization engine automatically writes the on-site measured coordinates back to the cloud BIM model, realizes the dynamic synchronization of the design model and the physical building, and provides full life cycle data support for project quality control.

[0053] It is obvious to a person skilled in the art that the application is not limited to the details of the exemplary embodiments described above, but can be implemented in other concrete forms without departing from the spirit or essential characteristics of the application. Therefore, the embodiments should be considered exemplary and non-limiting, and the scope of the application is defined by the appended claims rather than the above description, and all changes falling within the meaning and scope of the equivalent elements of the claims are intended to be included in the application. Any reference signs in the claims should not be considered as limiting the claims involved.

[0054] Furthermore, it should be understood that although the specification is described in terms of embodiments, not every embodiment includes every feature or implementation described herein. The specification can include implicit combinations of explicitly mentioned features and / or implicit combinations of implicitly mentioned features. Such combinations are also expressly included within the scope of the specification and an embodiment.

Claims

1. A BIM intelligent assembly system applied to node construction, characterized in that, Includes the following modules: Node-level BIM parametric model library: composed of LOD400 level node families, with embedded collision rules, rebar avoidance algorithms and tolerance limits within the families; Factory-side intelligent manufacturing module: It communicates with the node-level BIM parametric model library to drive the rebar bending robot, sleeve automatic welding machine and 3D printing positioning mold to manufacture node embedded parts with an accuracy of ≤0.5mm; The on-site intelligent assembly module includes a node perception layer composed of a reusable UWB+IMU+vision fusion sensing kit, which acquires the six-degree-of-freedom pose of the node rebar and sleeve in real time; an edge decision layer with an embedded lightweight Transformer model, which predicts the assemblability of the node based on the real-time pose and BIM theoretical coordinates and outputs the pose adjustment strategy; and a self-locking and self-adjusting execution layer integrated with the shape memory alloy locking pin and piezoelectric micro-displacement slide in the node connection plate, which automatically completes the rebar hole alignment, sleeve fine adjustment and temporary locking according to the pose adjustment strategy. The blockchain quality archive sub-module writes the entire process data of the node from material feeding in the factory to grouting completion into the consortium blockchain in the form of BIM-hash, realizing lifelong traceability of node-level quality.

2. The BIM intelligent assembly system applied to node construction according to claim 1, characterized in that, The node-level BIM parametric model includes an automatic node type classifier for automatically matching the optimal node family based on structural calculation results and a tolerance tolerance dynamic adjuster for real-time correction of manufacturing tolerances based on on-site environmental temperature and humidity and component age.

3. The BIM intelligent assembly system applied to node construction according to claim 2, characterized in that, The intelligent manufacturing module at the factory end includes a real-time visual closed-loop feedback unit for the end effector of the rebar bending robot, ensuring that the actual bending error is ≤0.3mm. The intelligent manufacturing module at the factory end also includes a laser weld seam tracking unit for the automatic sleeve welding machine, ensuring that the sleeve coaxiality is ≤0.2mm.

4. A BIM intelligent assembly system applied to node construction according to claim 3, characterized in that, The UWB+IMU+vision fusion sensing kit in the node perception layer adopts a magnetic quick-release structure; the Transformer model of the edge decision layer is quantized and pruned, with a model size ≤10MB, inference latency ≤100ms, and runs in the edge computing box inside the crane cab; the shape memory alloy locking pin of the self-locking and self-adjusting execution layer has a driving current of 0.8-1.2A, a locking force ≥3kN, and a response time ≤200ms; the piezoelectric micro-displacement slide has a stroke of ±2mm, a closed-loop resolution ≤0.5μm, and a maximum load ≥500N.

5. A BIM intelligent assembly system applied to node construction according to claim 4, characterized in that, The blockchain quality archive submodule adopts BIM-hash specifically by combining the GUID, manufacturing batch number, on-site sensor data hash value, and timestamp of the node BIM model object to generate a Merkle root and writing it into the consortium blockchain.

6. A BIM intelligent assembly system applied to node construction according to claim 5, characterized in that, It also includes a digital twin synchronization engine, which automatically writes back the measured coordinates on site to the cloud BIM model after the node assembly is completed, updating the deviation between the model coordinates and the entity coordinates to ≤0.1mm.

7. A BIM intelligent assembly system applied to node construction according to claim 6, characterized in that, Includes the following steps: S1. In the BIM platform, call the node-level parametric family to complete the collision check and rebar avoidance optimization, and generate manufacturing instructions; S2. The factory completes the intelligent manufacturing of node embedded parts according to the manufacturing instructions, and embeds RFID identification tags in the embedded parts; S3. Before on-site hoisting, BIM node instances are automatically matched via RFID identification and distributed to the edge decision-making layer; S4. During the hoisting process, the node perception layer collects the position and pose of the rebar / sleeve in real time, and the edge decision layer outputs the pose adjustment strategy. S5. The self-locking and self-adjusting execution layer completes the fine-tuning and temporary locking of the reinforcing bars to the holes and sleeves according to the strategy. S6. After grouting is completed, the blockchain quality archive sub-module will upload all process data to the blockchain, and the digital twin synchronization engine will update the BIM model.

8. A BIM intelligent assembly system applied to node construction according to claim 7, characterized in that, In step S4, when the predicted assemblability score is >95%, the crane micro-motion mode is automatically triggered, and millimeter-level translation or rotation commands are sent to the crane PLC through the edge computing box.

9. A BIM intelligent assembly system applied to node construction according to claim 8, characterized in that, If the real-time monitoring value of the grouting intensity at the node in step S6 is lower than 90% of the design value, the next process will be automatically suspended through the blockchain smart contract and an early warning will be pushed to the supervision APP.

10. A BIM intelligent assembly system applied to node construction according to claim 9, characterized in that, On-site data is uploaded to the cloud via a 5G network, with an end-to-end latency of ≤20ms.

Images