Glass curtain wall construction operation and maintenance safety intelligent monitoring and early warning system based on integration of BIM and Internet of Things

Abstract

The invention relates to the technical field of curtain wall construction, and discloses a BIM and Internet of Things integration-based glass curtain wall construction operation and maintenance safety intelligent monitoring and early warning system, which comprises a sensing layer, a stress sensor array deployed on a curtain wall unit, a three-axis vibration sensor, a multispectral environment sensor, a high-definition track camera and a UWB positioning module, the edge layer comprises a Kalman filtering unit and a mixed compression unit based on LZW-Huffman, wherein the Kalman filtering unit and the mixed compression unit are integrated on a sensor terminal; the system comprises a cloud platform layer, a dynamic BIM engine, a multi-source data fusion center, a structure behavior knowledge graph database, a hierarchical decision maker and a feedback learning module. And an early warning execution layer. According to the invention, high-precision and low-delay data acquisition is realized through the sensing layer multi-source sensor array and the edge layer real-time filtering compression unit, the problem of data islands in the prior art is effectively solved, active prediction of failure conduction is realized through a real-time deduction engine and a weight adaptive module of a structure knowledge graph, and the reliability of the system is improved. The problem of weak risk prediction in the prior art is solved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of curtain wall construction, in particular to a glass curtain wall construction operation and maintenance safety intelligent monitoring and early warning system based on BIM and Internet of Things integration. BACKGROUND

[0002] With the popularity of super high-rise buildings, glass curtain walls are widely used due to lighting and aesthetic advantages, but their construction and operation and maintenance safety problems are increasingly prominent. The current mainstream monitoring technologies include: sensor discrete monitoring, BIM static model and manual inspection system.

[0003] The existing technology has the following bottlenecks:

[0004] The hoisting positioning data (UWB) in the construction phase, the stress monitoring data and the crack detection data in the operation and maintenance period belong to independent systems, and there is a lack of a full life cycle data fusion platform;

[0005] Only fixed threshold alarm (such as stress > 50MPa triggering alarm) cannot identify the conduction chain of "loose bolt -> stress redistribution -> sealing failure -> glass falling off";

[0006] It cannot locate the accurate position of the damage in the three-dimensional space, resulting in poor pertinence of the maintenance scheme;

[0007] The static rule base cannot adapt to different curtain wall structure topological characteristics (such as point-supported vs. framed), and has no self-learning ability.

[0008] In view of this, we propose a glass curtain wall construction operation and maintenance safety intelligent monitoring and early warning system based on BIM and Internet of Things integration. SUMMARY

[0009] The purpose of the present application is to provide a glass curtain wall construction operation and maintenance safety intelligent monitoring and early warning system based on BIM and Internet of Things integration to solve the problems raised in the background art.

[0010] To achieve the above purpose, the present application provides the following technical solutions:

[0011] The glass curtain wall construction operation and maintenance safety intelligent monitoring and early warning system based on BIM and Internet of Things integration comprises:

[0012] The perception layer is deployed in the stress sensor array, three-axis vibration sensor, multi-spectral environment sensor, high-definition track camera and UWB positioning module of the curtain wall unit;

[0013] The edge layer is integrated in the Kalman filter unit and the mixed compression unit based on LZW-Huffman of the sensor terminal;

[0014] Cloud platform layer, dynamic BIM engine, multi-source data fusion center, structural behavior knowledge graph library, hierarchical decision maker, feedback learning module;

[0015] The early warning execution layer includes AR visualization terminals and multi-level sound and light alarm matrices.

[0016] Preferably, the dynamic BIM engine includes:

[0017] Component topology management submodule: splits the curtain wall BIM model into N independent units according to function, and constructs a mapping function between unit ID and sensor physical address;

[0018] State-driven submodule: parses sensor data in real time and updates unit state attributes.

[0019] Preferably, the multi-source data fusion center includes:

[0020] Spatiotemporal calibration submodule: aligns UWB coordinates with BIM space through quaternion rotation transformation;

[0021] Strain field construction submodule: generates a three-dimensional strain field by fusing stress array and vibration sensor data;

[0022] Damage identification submodule: uses the ResNet-34 convolutional network to process track camera images and output damage index.

[0023] Preferably, the structure-behavior knowledge graph library comprises:

[0024] Failure graph database: stores the node relationship topology of historical accidents;

[0025] Real-time deduction engine: Activates the conduction chain based on damage index and strain field;

[0026] Weight adaptation submodule: updates edge confidence according to maintenance records.

[0027] Preferably, the grading decision maker performs:

[0028] Risk quantification submodule: calculate unit risk value;

[0029] Key node marking submodule: assign key attributes to the conduction chain;

[0030] Early warning drive submodule: maps key attributes to BIM unit ID and triggers terminal highlighting.

[0031] Preferably, the feedback learning module performs:

[0032] Labeling interface: receiving the user's revised labels for warning results;

[0033] Model update module: Dynamically optimize the damage identification network.

[0034] Preferably, the AR visualization terminal comprises:

[0035] Dynamic rendering engine: superimposed display of three layers;

[0036] Spatial positioning superimposer: BIM model and physical curtain wall centimeter-level alignment through ARKit.

[0037] Preferably, the multi-level sound and light alarm matrix comprises:

[0038] First level: trigger terminal risk layer pulse;

[0039] Second level: start sound and light alarm;

[0040] Third level: activate full building stroboscopic.

[0041] By the above technical solutions, the present application provides a glass curtain wall construction operation and maintenance safety intelligent monitoring and early warning system based on BIM and Internet of Things integration. At least the following beneficial effects are achieved:

[0042] Through the perception layer multi-source sensor array and the edge layer real-time filtering and compression unit, high-precision and low-latency data acquisition is realized, effectively solving the data island problem existing in the prior art; through the component topology management and state driving module of the dynamic BIM engine, three-dimensional centimeter-level positioning of damage is realized, effectively solving the positioning accuracy problem existing in the prior art; through the time and space calibration and damage identification of the multi-source data fusion center, multi-modal data intelligent fusion is realized, solving the extensive diagnosis problem existing in the prior art; through the real-time deduction engine and weight adaptive module of the structure knowledge graph, failure conduction active prediction is realized, solving the weak risk prediction problem existing in the prior art; through the risk quantification module and key node marking of the hierarchical decision maker, precise hierarchical early warning is realized, solving the static threshold limitation problem existing in the prior art; through the labeling interface and model updater of the feedback learning module, system self-evolution is realized, solving the lack of intelligence problem existing in the prior art; through the three-layer superimposed rendering and ARKit positioning of the AR visualization terminal, risk space visualization is realized, solving the unclear three-dimensional positioning problem existing in the prior art; through the three-level triggering mechanism of the multi-level sound and light alarm matrix, emergency response automation is realized, solving the response lag problem existing in the prior art. BRIEF DESCRIPTION OF DRAWINGS

[0043] The accompanying drawings incorporated in and forming a part of the specification illustrate further aspects of the present application, and together with the description serve to explain the principles of the application:

[0044] Figure 1 FIG. 1 is a structural schematic diagram of the glass curtain wall construction operation and maintenance safety intelligent monitoring and early warning system based on BIM and Internet of Things integration of the present application. DETAILED DESCRIPTION

[0045] The technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all the embodiments of the present application. Based on the embodiments in the present application, all other embodiments obtained by a person of ordinary skill in the art without creative effort belong to the scope of protection of the present application.

[0046] Please refer to Figure 1 as shown, Figure 1 It is a structural schematic diagram of the glass curtain wall construction operation safety intelligent monitoring and early warning system based on BIM and Internet of Things integration. The present application provides a glass curtain wall construction operation safety intelligent monitoring and early warning system based on BIM and Internet of Things integration, which comprises:

[0047] The perception layer 101 is deployed in the stress sensor array, three-axis vibration sensor, multi-spectral environment sensor, high-definition track camera and UWB positioning module of the curtain wall unit.

[0048] The edge layer 102 is integrated in the Kalman filter unit and the mixed compression unit based on LZW-Huffman of the sensor terminal.

[0049] The cloud platform layer 103 is a dynamic BIM engine, a multi-source data fusion center, a structure behavior knowledge graph library, a hierarchical decision maker and a feedback learning module.

[0050] The early warning execution layer 104 comprises an AR visualization terminal and a multi-stage sound and light alarm matrix.

[0051] It should be noted that the perception layer configuration comprises a four-corner pre-embedded FBG stress sensor array of each curtain wall unit (1.5m x 4m), with a range of 0-100MPa; the UWB positioning module is embedded in the aluminum frame, with a transmission power of 10dBm and a positioning accuracy of ±3mm.

[0052] The edge layer parameters include a Kalman filter cutoff frequency of 50Hz to suppress hoisting vibration noise; the LZW-Huffman mixed compression has a stress data compression ratio of 8:1, and the image ROI region is lossless compressed.

[0053] As an option, the dynamic BIM engine comprises:

[0054] The component topology management sub-module: the curtain wall BIM model is divided into N independent units according to function, and a mapping function of unit ID and sensor physical address is constructed.

[0055] The state driving sub-module: real-time analysis of sensor data and updating of unit state attributes.

[0056] It should be noted that the component topology management submodule establishes a bidirectional index of unit ID and sensor address through a SHA-256 hash function, ensuring that construction and operation data are uniformly collected; the state driving submodule dynamically adjusts the threshold value (σ_max = 0.8 x σ_yield) based on the material yield strength, and the displacement vector coordinate transformation adopts the Lie group SE(3) theory, with the bias set as [0.35, -0.28, 0]^T mm;

[0057] The component topology management submodule adopts spatial hash mapping technology to split the curtain wall BIM model into independent units, and then generates a binding relationship between unit ID and sensor physical address through the SHA-256 algorithm. Each unit ID is associated with a group of sensor addresses (including stress sensors, UWB positioning modules, etc.). The hash table capacity is set to 1024 slots, and the actual conflict rate is less than 0.3%. When new sensor nodes are added, the system automatically performs consistent hash redistribution, and the expansion delay is controlled within 15 milliseconds.

[0058] The physical quantity conversion rule of the state driving submodule is:

[0059] When the stress value is mapped to the hue, when the stress is less than or equal to 80% of the material yield strength, the hue value changes from blue (240°) to yellow (60°) in a linear gradient; when the threshold is exceeded, the hue non-linearly jumps to the red warning zone (<48°);

[0060] The displacement calibration adopts a three-dimensional space compensation mechanism: the preset installation bias compensation amounts X / Y / Z axes are 0.02 radians, -0.01 radians, and 0.03 radians, respectively, and the translation compensation amounts are set to 0.35 mm, -0.28 mm, and 0.12 mm.

[0061] As an option, the multi-source data fusion center includes:

[0062] The space-time calibration submodule aligns the UWB coordinates and the BIM space through quaternion rotation transformation;

[0063] The strain field construction submodule generates a three-dimensional strain field by fusing stress array and vibration sensor data;

[0064] The damage identification submodule uses a ResNet-34 convolutional network to process track camera images and outputs a damage index.

[0065] It should be noted that the space-time calibration submodule introduces gravity acceleration compensation, and the gravity vector and curtain wall normal vector are fused in the calculation of the quaternion rotation matrix, solving the coordinate system drift caused by wind vibration of high-rise buildings; the strain field construction submodule uses a radial basis function interpolation (RBF) algorithm to fuse discrete sensor data with a cubic basis function φ(r) = r^3.

[0066] The spatio-temporal calibration sub-module integrates IMU inertial measurement data, compensates for high-level wind vibration effects through quaternion rotation, and sets the core conversion matrix as a 3x3 rotation part (precision 0.15°RMS) and a 3x1 translation part (precision 0.28mmRMS);

[0067] The strain field construction sub-module uses an improved radial basis interpolation algorithm: generating a Delaunay triangular network with each sensor as a node; based on the gradient correction of the shape function, the interpolation weight is modified to improve the spatial continuity of the strain field by 40%.

[0068] As an option, the structural behavior knowledge graph library includes:

[0069] Failure graph database: stores the node relationship topology of historical accidents;

[0070] Real-time inference engine: activates the conduction chain based on the damage index and strain field;

[0071] Weight adaptive sub-module: updates the edge confidence according to maintenance records.

[0072] It should be noted that the real-time inference engine sets a time decay condition: only when the damage duration is >5 minutes, the conduction chain is activated to prevent false triggering of transient interference; the graph database uses Neo4j to store failure topology, and supports Cypher language query.

[0073] Dual conditions for the real-time inference engine to activate the conduction chain: damage index α>0.4 and strain gradient exceeds the allowed value by 20%, and abnormal state lasts more than 5 minutes (to prevent transient interference);

[0074] The graph database selects the Neo4j architecture, and retrieves high-risk conduction paths through the Cypher query language (example: search for three-level conduction chains with weight>0.7).

[0075] As an option, the hierarchical decision maker performs:

[0076] Risk quantification sub-module: calculates the risk value of the unit;

[0077] Key node marking sub-module: assigns key attributes to the conduction chain;

[0078] Early warning driving sub-module: maps the key attributes to the BIM unit ID to trigger terminal highlighting.

[0079] It should be noted that the risk quantification submodule dynamically adjusts the weight coefficients for different types of curtain walls: [0.6, 0.3, 0.1]^T for unit curtain walls, and [0.5, 0.4, 0.1]^T for point-supported curtain walls; the key node screening is based on the graph theory betweenness centrality C_B(v), and only when the node betweenness is greater than 2 / (n(n-1)) is it marked as a key node (n is the total number of nodes); the key node screening is based on the graph theory betweenness centrality index, and only when the node is located in more than 20% of the shortest transmission path is it marked as a key point.

[0080] As an option, the feedback learning module performs:

[0081] Label interface: receives user correction labels for early warning results;

[0082] Model update module: dynamically optimizes damage identification network.

[0083] It should be noted that the model updater adopts a transfer learning mechanism: the bottom convolutional layers of ResNet-34 are frozen, only the top layer parameters are fine-tuned, and a regularization term (γ=0.01) is added to the loss function to prevent catastrophic forgetting; user correction labels δ trigger graph edge weight updates, and the update amount is multiplied by a sigmoid function decay factor.

[0084] The model updater adopts a transfer learning strategy:

[0085] Freeze the bottom convolutional layers of the ResNet-34 network;

[0086] Fine-tune only the top fully connected layer weights;

[0087] The learning rate is decayed from 0.01 to 0.0001 according to the cosine annealing curve;

[0088] The knowledge graph optimization includes user correction labels triggering edge weight updates, and the update amplitude decays exponentially with the length of the transmission chain.

[0089] As an option, the AR visualization terminal includes:

[0090] Dynamic rendering engine: superimposes three layers of images;

[0091] Spatial positioning superimposer: achieves centimeter-level alignment of BIM models and physical curtain walls through ARKit.

[0092] It should be noted that the dynamic rendering engine uses Alpha blending technology to achieve three-layer image fusion, with risk layer transparency α_risk=0.7 and heat layer α_heat=0.4; the spatial positioning superimposer fuses ARKit and UWB data, and the position calculation uses weighted calculation; the spatial positioning fuses ARKit visual positioning and UWB signals (weight ratio 7:3), and the jitter error after Kalman filtering is ≤0.7mm.

[0093] Optionally, the multi-stage sound-light alarm matrix comprises:

[0094] First stage: trigger terminal risk layer pulse;

[0095] Second stage: start sound-light alarm;

[0096] Third stage: activate full building stroboscopic.

[0097] It should be noted that the second stage alarm frequency response control adopts a segmented function: when the risk value R < 0.5, it is fixed at 1 kHz, and when R ≥ 0.5, it is adaptively adjusted according to f = 2R^2 + 1 (kHz), and the light pulse frequency is positively correlated with the risk value; the red light stroboscopic of the third stage alarm uses Manchester coding to transmit position information, and a 16-bit binary sequence coding unit ID is used.

[0098] The present application realizes high-precision and low-delay data acquisition through the perception layer multi-source sensor array and the edge layer real-time filtering and compression unit, effectively solving the data island problem existing in the prior art; through the component topology management and state driving module of the dynamic BIM engine, the three-dimensional centimeter-level positioning of damage is realized, effectively solving the problem of insufficient positioning accuracy in the prior art; through the space-time calibration and damage identification of the multi-source data fusion center, the intelligent fusion of multi-modal data is realized, solving the problem of extensive diagnosis in the prior art; through the real-time deduction engine and weight adaptive module of the structure knowledge graph, the active prediction of failure transmission is realized, solving the problem of weak risk prediction in the prior art; through the risk quantification module and key node marking of the hierarchical decision maker, precise hierarchical early warning is realized, solving the problem of static threshold limitation in the prior art; through the annotation interface and model updater of the feedback learning module, system self-evolution is realized, solving the problem of lack of intelligence in the prior art; through the three-layer superposition rendering of the AR visualization terminal and the ARKit positioning, risk space visualization is realized, solving the problem of unclear three-dimensional positioning in the prior art; through the three-stage triggering mechanism of the multi-stage sound-light alarm matrix, emergency response automation is realized, solving the problem of response lag in the prior art.

[0099] It should be noted that in this document, the terms such as first and second are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "contain" or any other variants thereof are intended to cover non-exclusive inclusion, so that the process, method, article or equipment including a series of elements not only includes those elements, but also includes other elements not explicitly listed or inherent to such process, method, article or equipment.

[0100] While embodiments of the application have been shown and described, it is to be understood that the embodiments described are merely exemplary of the principles and application of the present application. Numerous modifications and adaptions can be effected without departing from the spirit and scope of the present application, which is not limited to the exact construction and arrangement described. It is intended, therefore, to cover all modifications and adaptions that fall within the scope of the claims and their equivalents.

Claims

1. The intelligent monitoring and early warning system for glass curtain wall construction and operation safety based on the integration of BIM and the Internet of Things is characterized by: include: The perception layer includes stress sensor arrays, three-axis vibration sensors, multi-spectral environmental sensors, high-definition track cameras, and UWB positioning modules deployed on curtain wall units. Edge layer, Kalman filter unit and LZW-Huffman based hybrid compression unit integrated in the sensor terminal; Cloud platform layer, dynamic BIM engine, multi-source data fusion center, structural behavior knowledge graph library, hierarchical decision maker, feedback learning module; The early warning execution layer includes AR visualization terminals and multi-level sound and light alarm matrices.

2. The glass curtain wall construction and operation safety intelligent monitoring and early warning system based on BIM and Internet of Things integration according to claim 1 is characterized in that: The dynamic BIM engine includes: Component topology management submodule: splits the curtain wall BIM model into N independent units according to function, and constructs a mapping function between unit ID and sensor physical address; State-driven submodule: parses sensor data in real time and updates unit state attributes.

3. The glass curtain wall construction and operation safety intelligent monitoring and early warning system based on BIM and Internet of Things integration according to claim 1 is characterized in that: The multi-source data fusion center includes: Spatiotemporal calibration submodule: aligns UWB coordinates with BIM space through quaternion rotation transformation; Strain field construction submodule: generates a three-dimensional strain field by fusing stress array and vibration sensor data; Damage identification submodule: uses the ResNet-34 convolutional network to process track camera images and output damage index.

4. The glass curtain wall construction and operation safety intelligent monitoring and early warning system based on BIM and Internet of Things integration according to claim 1 is characterized in that: The structural behavior knowledge graph library includes: Failure graph database: stores the node relationship topology of historical accidents; Real-time deduction engine: Activates the conduction chain based on damage index and strain field; Weight adaptation submodule: updates edge confidence according to maintenance records.

5. The glass curtain wall construction and operation safety intelligent monitoring and early warning system based on BIM and Internet of Things integration according to claim 1 is characterized in that: The grading decider performs: Risk quantification submodule: calculate unit risk value; Key node marking submodule: assign key attributes to the conduction chain; Early warning drive submodule: maps key attributes to BIM unit ID and triggers terminal highlighting.

6. The glass curtain wall construction and operation safety intelligent monitoring and early warning system based on BIM and Internet of Things integration according to claim 1 is characterized in that: The feedback learning module performs: Labeling interface: receiving the user's revised labels for warning results; Model update module: Dynamically optimize the damage identification network.

7. The glass curtain wall construction and operation safety intelligent monitoring and early warning system based on BIM and Internet of Things integration according to claim 1 is characterized in that: The AR visualization terminal includes: Dynamic rendering engine: superimpose and display three layers; Spatial Positioning Overlay: Use ARKit to achieve centimeter-level alignment between BIM models and physical curtain walls.

8. The glass curtain wall construction and operation safety intelligent monitoring and early warning system based on BIM and Internet of Things integration according to claim 1 is characterized in that: The multi-stage sound and light alarm matrix includes: Level 1: triggering the terminal risk layer pulse; Level 2: activate the sound and light alarm; Level 3: Activate the whole building strobe.

Images