Internet-of-things-perceived visual management method for on-site construction of project to be built

Abstract

The invention discloses a visual management method for on-site construction of a construction project perceived by the Internet of Things, and belongs to the technical field of construction engineering management. Comprising the following steps: constructing a digital twin base model; deploying a multi-mode Internet of Things sensing network and performing data semantic mapping; creating and driving a dynamic digital twinborn body in real time; constructing and solving a construction element coupling risk potential field model; a transmissive visual interaction is provided. Dynamic elements of a construction site are abstracted into'potential points' carrying instantaneous risk potential, and interaction force among the potential points and composite risk potential of any point in space are calculated in real time through a mathematical formula defining risk gravitation and safety repulsive force. By solving the model, the method can quantify and visualize a dynamic composite risk generated by superposition of a plurality of risk sources, realizes predictive early warning from an event result to a risk cause, and can generate a navigation type optimal risk avoiding path instruction.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of construction engineering management, in particular to a construction site visual management method applied to an agent construction project management mode, which integrates Internet of Things (IoT), Digital Twin and Artificial Intelligence (AI) technologies. BACKGROUND

[0002] The agent construction project management mode (Agent Construction Model) refers to that a government or a principal party (a principal party) selects a professional project management unit (an agent party) through bidding and other ways, and the agent party is responsible for the whole process management of investment management, construction organization and implementation of a project until project completion acceptance and handover for use. This mode is increasingly common in large public buildings and infrastructure construction.

[0003] Currently, the construction site management of the agent construction project mainly relies on traditional management means, for example: Manual inspection and reporting: supervising engineers and site management personnel manually record construction progress, quality and safety issues through regular or irregular patrol, and report to the superior through meetings, paper reports or simple electronic spreadsheets. This way is highly subjective, information transmission is lagging, data granularity is coarse, and it is difficult to achieve real-time, comprehensive and accurate control of complex sites.

[0004] Video monitoring: video monitoring cameras (CCTV) are generally installed in construction sites, but are mostly used for post-tracing or simple remote viewing, lacking intelligent analysis capability of video data. Management personnel need to spend a lot of time watching long video recordings, and cannot actively find problems and warn risks.

[0005] BIM (Building Information Modeling) application: although BIM technology realizes three-dimensional visualization of building design and construction, in the construction phase, the BIM model is often disconnected from the actual construction site, becoming a static and isolated "digital sandbox". The dynamic information such as actual construction progress, resource status and environmental changes cannot be fed back to the BIM model in real time, resulting in a serious separation between "digital" and "physical" worlds, and the guiding and management value of BIM is greatly discounted.

[0006] The existing technology has the following main defects: Data islandization: progress, quality, safety, material, personnel and other data on site are scattered in different systems or paper records, which cannot be effectively associated and integrated, and it is difficult to form a global management view.

[0007] Management passivity: the management mode is mainly based on "after-the-fact remediation", lacking the ability of "pre-early warning" and "in-process control" based on real-time data. For example, problems can only be found after the report of substandard concrete strength, and real-time intervention cannot be made during pouring or curing.

[0008] Information asymmetry: the information transmission chain between the principal, the construction agent, the construction party, the supervision party and the like is long and inefficient, and information distortion and decision delay are prone to occur, especially in the construction agent mode, the principal is difficult to penetrate to know the real and real-time site situation.

[0009] Low degree of visualization: the existing visualization is mostly limited to static BIM models or isolated video pictures, and cannot real-time fuse the multi-dimensional information of dynamic people, machines, materials, methods and environment with the space model to form a dynamic, fresh and interactive "digital twin construction site".

[0010] Therefore, a new construction management method for construction agent projects is urgently needed, which can break data silos, realize active early warning, enhance the efficiency of multi-party collaboration, and provide deep visualization insight. SUMMARY

[0011] This section aims to summarize some aspects of the embodiments of the present application and briefly introduce some preferred embodiments. Some simplifications or omissions may be made in this section and the abstract and title of the specification to avoid obscuring the purpose of this section, abstract and title, and such simplifications or omissions cannot be used to limit the scope of the present application.

[0012] In view of the above existing problems, the present application is proposed.

[0013] The present application aims to solve the problems in the existing construction management method, such as static and isolated risk identification, inability to quantify the complex risks generated by the coupling of dynamic elements, and late warning. The present application provides a new method for real-time calculation, prediction and visualization of dynamic complex risks in construction sites, thereby realizing active and predictive safety and progress management from "event-driven" to "field-driven".

[0014] To solve the above technical problems, the present application provides the following technical solution: an Internet of Things sensing construction agent project site construction visualization management method, comprising the following steps: a) constructing a digital twin base model comprising a building information model, geographic information system data and a site three-dimensional model; b) deploying an Internet of Things sensing network at the construction site, acquiring real-time sensing data representing the state of personnel, machinery, materials and environment, and establishing a semantic mapping relationship between the real-time sensing data and the model units in the digital twin base model; c) based on the mapping relationship, driving the digital twin base model using the real-time perception data to generate a real-time updated dynamic digital twin; d) performing a risk quantification step comprising: d1) abstracting dynamic construction elements in the dynamic digital twin as potential points, and for each potential point P i determining its instantaneous risk potential V i (t) and real-time position vector ; d2) according to a preset coupling type coefficient a ij and distance influence attenuation index n, calculating the risk force vector between any two potential points P i and P j ; wherein: ; d3) generating a warning instruction based on the calculated risk force vector; e) superimposing information generated by the risk quantification step on the visualization interface of the dynamic digital twin.

[0015] As a preferred scheme of the Internet of Things perception-based construction project site construction visualization management method of the present application, in the step d1), the instantaneous risk potential V i (t) is calculated by the model V i (t) = V int,i *(1+δ i (t)); wherein, V int,i is a preset inherent risk potential, and δ i (t) is a state disturbance coefficient determined according to the real-time perception data.

[0016] As a preferred scheme of the Internet of Things perception-based construction project site construction visualization management method of the present application, in the step d3), specifically comprising: calculating the total risk force received by any potential point P k ; ; when the modulus of the total risk force exceeds a first preset threshold, generating a warning instruction containing the identification of the potential point.

[0017] As a preferred scheme of the Internet of Things perception-based construction project site construction visualization management method of the present application, the warning instruction further comprises directional vector data determined according to the opposite direction of the total risk force. ​

[0018] As a preferred scheme of the Internet of Things aware construction site visualization management method of the agent project, wherein: the risk quantification step d) further comprises: According to the model, the composite risk potential of the predetermined point set or region in the construction site space is calculated . When the composite risk potential of any region exceeds the second preset threshold, a warning instruction containing the location information of the region is generated; Wherein: .

[0019] As a preferred scheme of the Internet of Things aware construction site visualization management method of the agent project, wherein: according to the calculation result of the composite risk potential A three-dimensional risk heat map is rendered in the three-dimensional space of the dynamic digital twin.

[0020] As a preferred scheme of the Internet of Things aware construction site visualization management method of the agent project, wherein: the step e) specifically comprises: in response to the selection operation of the user on any potential point in the dynamic digital twin, one or more risk action force vectors .

[0021] As a preferred scheme of the Internet of Things aware construction site visualization management method of the agent project, wherein: the method further comprises a virtual-real deviation analysis step, which comprises: Obtain the real three-dimensional point cloud model of the scene; Register the point cloud model with the corresponding part in the digital twin base model and calculate the geometric deviation; Highlight the area with deviation value exceeding the third preset threshold in the dynamic digital twin.

[0022] As a preferred scheme of the Internet of Things aware construction site visualization management method of the agent project, wherein: the method further comprises a risk trend prediction step, which comprises: Collect the composite risk potential The numerical value of a certain key region in a continuous time period forms time series data; Input the time series data into a time series prediction model to output the predicted value of the future composite risk potential of the region; When the predicted value exceeds the second preset threshold, a warning instruction is generated.

[0023] The present application provides an Internet of Things aware construction site visualization management method of the agent project, which has the following advantages: 1. The CE-CRPF model, for the first time, transforms abstract, discrete construction risks into calculable, superimposed, and continuously distributed physical "fields." This makes it possible to accurately quantify the composite risks generated by the coupling of multiple dynamic risk sources, representing a major breakthrough in construction safety management theory and methods.

[0024] 2. Based on the calculation of risk forces and risk gradients, the system can issue early warnings by analyzing the dynamic process of "approaching" risk events before they actually occur. It provides personnel with "navigation-style" proactive risk avoidance paths, achieving a qualitative leap from passive alarm to proactive guidance, and greatly improving the effectiveness of early warnings. Detailed Implementation

[0025] To make the above-mentioned objects, features, and advantages of the present invention more apparent and understandable, specific embodiments of the present invention will be described in detail below. Obviously, the described embodiments are only a part of the embodiments of the present invention, and not all of them. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort should fall within the protection scope of the present invention.

[0026] This invention aims to address the problems of static and isolated risk identification in existing construction management methods, the inability to quantify complex risks arising from the coupling of dynamic factors, and delayed early warnings. This invention provides a new method for proactive and predictive safety and schedule management that can calculate, predict, and visualize dynamic complex risks on construction sites in real time, thereby shifting from "event-driven" to "site-driven" approaches.

[0027] Therefore, this invention provides an IoT-based method for visual management of on-site construction in construction management projects, comprising the following steps: S1: Construct a project-level digital twin base model.

[0028] During the project preparation phase, Building Information Modeling (BIM), Geographic Information System (GIS) data, and a detailed 3D site model are integrated to construct a digital twin base model (DTB-Model) that precisely corresponds to the physical construction site in terms of geometric, spatial, and topological dimensions. This model includes not only the building structure but also temporary facilities, pipeline routes, pre-designed parking areas for large machinery, and material storage yards.

[0029] S2: Deploy a multimodal IoT sensing network and perform semantic mapping of data.

[0030] Deploy an IoT sensing network consisting of various types of sensors at the construction site and establish a semantic mapping relationship between sensor data and model units in the digital twin base model.

[0031] S21: Perception of network deployment: Personnel perception: Wear integrated UWB (Ultra Wide Band) / Bluetooth AOA (Angle of Arrival) high-precision positioning modules, vital sign monitoring (heart rate, blood oxygen), and SOS one-key alarm functions on the smart safety helmet or work card of site managers and special operators.

[0032] Mechanical perception: Install GPS / Beidou positioning modules, working state sensors (such as hook load, angle, oil consumption, engine speed sensors), and anti-collision millimeter wave radars on large machinery such as tower cranes, excavators, and concrete pump trucks.

[0033] Material perception: Attach RFID (Radio Frequency Identification) tags or two-dimensional codes to key materials such as steel bars and prefabricated components, and deploy fixed readers at warehouse entrances and exits, processing areas, and hoisting points to achieve full life cycle tracking of materials.

[0034] Environmental perception: Deploy environmental and structural safety monitoring stations integrated with temperature and humidity, PM2.5 / PM10, noise, wind speed sensors, and structural stress / settlement monitors at key locations on site (such as foundation pits, high-formwork areas, and dust operation areas).

[0035] Progress and quality perception: Deploy AI visual analysis cameras to use computer vision technology to identify construction processes (such as steel bar binding, formwork erection, and concrete pouring) in real time, and periodically scan the site using 3D laser scanners or unmanned aerial vehicle oblique photography to generate high-precision point cloud data.

[0036] S22: Data semantic mapping: Establish a real-time mapping rule library. Bind the data stream of each sensor or perception device to the unique model unit ID in the DTB-Model. For example, map the real-time position data {X, Y, Z} of the smart safety helmet numbered WB001 to the virtual human model representing the worker; map the RFID tag state (such as "in warehouse" or "hoisted") of the steel bar numbered RC005 to the corresponding steel member attribute in the BIM model; map the "2F area A axis column is pouring concrete" event identified by the AI camera to the state update of the column model unit.

[0037] S3: Create and drive a dynamic digital twin (DDT) in real time.

[0038] Based on the digital twin base model of step S1 and the real-time semantic data stream of step S2, a dynamic digital twin (DDT) is created and continuously updated through a data fusion and state synchronization engine. The engine performs the following operations: S31: Data cleaning and fusion: denoising, format conversion and spatio-temporal alignment of the received multi-source heterogeneous data.

[0039] S32: State synchronization: real-time update of the fused data into the attributes of the corresponding model units in DDT according to the semantic mapping rules. This enables DDT to reproduce the dynamic changes of the physical construction site in real time and realistically, including personnel position and state, mechanical operation trajectory and load, material flow path, construction process progress, and environmental parameter changes.

[0040] S33: Virtual-real deviation analysis: automatic registration and deviation analysis of the on-site real scene three-dimensional point cloud model generated by the unmanned aerial vehicle / laser scanner and the contemporaneous planned BIM model, generating a visual progress deviation report (red for lag, green for advance) and a quality deviation report (such as flatness, verticality out-of-tolerance area highlighting).

[0041] S4: Construction element coupling risk potential field model (CE-CRPF) is constructed and solved to realize dynamic risk quantification and predictive early warning.

[0042] S41: "Potential point" abstraction and parameterization of risk elements: In the dynamic digital twin (DDT), each independent, movable, and state-variable construction element (such as workers, large machinery, and hazardous source areas) is abstracted as a "potential point" (Potential Point). Each potential point P i is assigned a set of dynamic parameters: Intrinsic Risk Potential (V int,i ): represents the risk level of the element itself, independent of the outside world. This is a normalized base value.

[0043] For example: the V int of an ordinary worker may be 0.1; the V int of a special operations worker performing a high-risk task (such as high-altitude welding) will dynamically increase to 0.8; the V int of a static hazard source (such as a distribution box) is a constant value of 0.9.

[0044] State Disturbance Coefficient δ i (t): a function that changes with time, reflecting the dynamic influence of the current state of the element on its risk potential.

[0045] For example: for workers, δ i (t) may be related to their vital signs (abnormal heart rate), fatigue level (calculated according to continuous working time), and certificate status (certificate expiration).

[0046] For machinery, δi (t) possibly related to whether it is overloaded, whether it is in a maintenance cycle, oil / electricity level, etc.

[0047] Position vector : Real-time three-dimensional coordinates {x i ,y i ,z i} provided by the Internet of Things positioning system.

[0048] Therefore, at any time t, the instantaneous risk potential (Instantaneous Risk Potential, V i (t)) of a potential point P i is defined as: V i (t)=V int,i *(1+δ i (t)); S42: Define the interaction force between potential points - "risk attraction" and "safety repulsion": An interaction "risk force field" will be generated between any two potential points P i and P j . The size of the force is proportional to their risk potential and inversely proportional to the nth power of the distance between them. More importantly, the coupling type coefficient α ij is introduced to define the nature of the force: Risk attraction (Risk Attraction, α ij >0): When two elements approach each other, it increases the risk, and they exhibit attraction between them.

[0049] For example: workers approach running excavators; workers approach unguarded edges; areas under the boom and workers.

[0050] Safety repulsion (Safety Repulsion, α ij <0): When two elements approach each other, it reduces the risk or belongs to the compliant operation, and they exhibit repulsion or no force between them.

[0051] For example: hot work site and nearby fire extinguishers; welds to be inspected and certified inspectors.

[0052] No interaction (α ij =0): Two elements are irrelevant.

[0053] At time t, the risk force vector generated by potential point P j to P i is defined as:

[0054] Model parameters: :P j The risk force generated at P i is a vector pointing to the direction of risk increase.

[0055] G: risk field constant, a scaling factor to scale the result to a meaningful interval (e.g. 0-100).

[0056] α ij : coupling type coefficient, a preset matrix by expert system or historical data mining, defining the interaction between the i-th element and the j-th element (attractive / repulsive / irrelevant).

[0057] V i (t), V j (t): instantaneous risk potential of two potential points, as defined in S41.

[0058] : position vector of two potential points.

[0059] : Euclidean distance between two points.

[0060] n: distance influence decay index, usually between 1 and 3. For example, for high-altitude falling object risk, n may take a larger value, indicating that the risk decays rapidly with distance; for noise influence, n may take a smaller value.

[0061] : unit vector from P j to P i , indicating the direction of the force.

[0062] S43: Calculate the composite risk potential and risk gradient at any point in space: The composite risk potential of any point in the construction site space is the scalar field generated by the combined action of all N potential points at that point. It is equal to the algebraic sum of the risk potential generated by each potential point at that point:

[0063] where m is another spatial decay index.

[0064] More instructive is to calculate the risk potential field gradient :

[0065] The direction of this gradient vector points to the direction of the fastest risk growth at that point, and its modulus represents the severity of the risk at this point.

[0066] S44: Predictive early warning and proactive intervention based on risk potential field model: Global risk heat map visualization: In the dynamic digital twin (DDT), the composite risk potential field of the entire construction site is calculated and rendered in real time . High-risk areas are displayed in red, and low-risk areas are displayed in blue, forming a dynamic, three-dimensional "risk weather map" that provides managers with an intuitive understanding of the overall risk situation.

[0067] Individual risk vector early warning: Calculate the total force on an individual: For any worker (or mobile device) P k , calculate the total risk force it receives from all other potential points:

[0068] Predictive early warning: When ∣∣ ∣∣ exceeds the preset "alert threshold" (not the "collision threshold"), the system issues an early warning. This represents that the worker is in a "gravitational center" where the risk is rapidly accumulating, even if he has not yet entered any specific dangerous area.

[0069] Intelligent navigation-based risk avoidance: Early warning information is not just "Danger!", but "Move 3 meters behind you to significantly reduce risk." This direction is the opposite direction of the total risk force vector . The system provides the optimal "safe evacuation path" in the form of an arrow on the worker's AR glasses or mobile app.

[0070] Prediction of group risk evolution trend: By using time series analysis methods (such as ARIMA model) to analyze the historical changes of the risk potential of key areas (such as foundation pit operation surface), the risk evolution trend of the area in the next 10 minutes or half an hour can be predicted. If it is predicted that it will break through the danger threshold, an evacuation warning will be issued to all personnel in the area in advance.

[0071] S5: Multi-dimensional, multi-role penetrating visual interaction (on the basis of the original, add visual interaction of risk potential field).

[0072] Risk field visualization layer: On the DDT visualization platform, users can turn on or off the "risk potential field" layer with one click. In this layer, the traditional BIM model is semi-transparent, and the real-time risk heat map is superimposed and displayed.

[0073] Interactive risk exploration: Users can click on any point in the three-dimensional scene with the mouse, and the system will immediately display the composite risk potential value , risk gradient vector and list the top three risk sources that contribute the most to this point risk.

[0074] Individual Risk Force Field Analysis: Click on any worker in the scene, the system will center on that worker and visualize all the "Risk Attraction" and "Safety Repulsion" vector arrows he is subjected to, so that the manager can immediately know "who is threatening him, and who is protecting him".

[0075] Example One: 1. Scene Setup: Time: January 28, 2025, 14:30:00; Location: A local area of a construction project, Building 2, 3rd floor, A area, where work is being carried out.

[0076] Coordinate System: Establish a right-handed three-dimensional coordinate system with the southwest corner of the floor as the origin (0, 0, 0), unit in meters.

[0077] Key "Potential Point" elements in the scene: P1 (Worker Zhang San): An ordinary worker who is doing steel reinforcement binding.

[0078] P2 (Tower Crane Load): A bundle of steel reinforcement (about 1.5 tons) suspended below a running tower crane.

[0079] P3 (Edge Hole): An incomplete closed floor reserved hole, which is a static danger source.

[0080] P4 (Fire Extinguisher): A standard dry powder fire extinguisher placed in the designated location, which is a safety facility.

[0081] 2. Parameter definition and example data.

[0082] 2.1 Basic Constants and Attenuation Exponents: Risk Field Constant G: Set to 100. This is a scaling coefficient to amplify the calculation results for easy observation and threshold setting.

[0083] Distance Influence Attenuation Exponent n (used for force field calculation): Set to 2. This indicates that the risk force is inversely proportional to the square of the distance, which conforms to the attenuation law of similar "fields".

[0084] Space Attenuation Exponent m (used for potential field calculation): Set to 1.

[0085] 2.2 Potential Point (Potential Point) Parameters

[0086] Note: V intSet to negative value, is a clever handling of the model, so that it is in the subsequent calculation of the natural performance of the "reduce risk" effect.

[0087] 2.3 Coupling type coefficient matrix (α ij ): This is a symmetric matrix, which defines the interaction between different types of elements.

[0088]

[0089] Matrix interpretation: α 12 =α 21 =1.0: There is a strong risk between workers and the hanging object. Attraction.

[0090] α 13 =α 31 =1.2: The risk between workers and the hole is stronger (more dangerous than the hanging object).

[0091] α 23 =α 32 =0.5: The hanging object is close to the hole, which also has some risk (such as collision causing the hole guardrail to be damaged).

[0092] α 14 =0: The fire extinguisher itself does not produce direct attraction or repulsion to the worker (its role is reflected elsewhere, such as the fire operation scene).

[0093] 3. Calculation process: 3.1 Step one: Calculate the instantaneous risk potential V i (t) of each potential point: V i (t)=V int,i *(1+δ i (t)); V1(t) (worker Zhang San): 0.2⋅(1+0.5)=0.3; V2(t) (tower crane hanging object): 0.9⋅(1+0)=0.9; V3(t) (edge hole): 0.8⋅(1+0.1)=0.88; V4(t) (fire extinguisher): −0.5⋅(1+0)=−0.5; Analysis: Due to fatigue, the instantaneous risk potential of worker Zhang San rises from the base 0.2 to 0.3. The hole also rises slightly due to the guardrail problem.

[0094] 3.2 Step two: Calculate the risk force vector on worker Zhang San (P1):

[0095] 3.2.1 Calculate the force from the crane load P2

[0096]

[0097] 3.2.2 Calculate the force from the edge opening P3

[0098]

[0099] 3.2.3 Calculate the force from the fire extinguisher P4

[0100]

[0101] 3.3 Step Three: Calculate the total risk force on worker Zhang San

[0102]

[0103] 3.4 Step Four: Analysis and Warning

[0104] 3.4.1 Calculate the total risk force module length:

[0105] 3.4.2 Warning Judgment Assume the system has two thresholds set: Alert Threshold: 1.0; Danger Threshold: 2.0; The current calculated module length 1.333 > 1.0 (Alert Threshold), but < 2.0 (Danger Threshold).

[0106] The system triggers a "Alert" level warning! 3.4.3 Generate warning instructions and analysis: Warning information: "Worker Zhang San (Employee ID G0128), you are in a risk accumulation area, please pay attention to safety!" Risk traceability analysis (for managers): The total risk force module length is 1.333.

[0107] Main risk contribution sources: edge opening (force module length contribution ≈1.20 ≈1.20), crane load (force module length contribution ≈0.53 ≈0.53).

[0108] ​​​​Conclusion: Although worker Zhang San is still far away from both risk sources (5.14m and 7.16m respectively), no simple distance-based e-fence is triggered. But this model successfully quantifies the compound risk scenario of "approaching both risk sources simultaneously" by calculating the force vector superposition, and finds that the cumulative risk has reached the level that requires alertness.

[0109] Smart navigation-based risk avoidance instruction: Total risk force vector This vector points to the direction where the risk grows fastest.

[0110] The optimal safe evacuation direction is the opposite direction of this vector: (1.313, 0.074, 0.221).

[0111] Instruction content (pushed to AR glasses or APP): A three-dimensional arrow appears on Zhang San's visual interface, pointing in the direction of (1.313, 0.074, 0.221), with the text: "Suggest moving to your right rear!" 4. Summary of the implementation example: This calculation example clearly shows how the method of the invention works: Quantify implicit risks: Factors such as worker fatigue and facility defects are incorporated into the risk calculation through the parameter δ i (t), making the risk assessment more comprehensive.

[0112] Identify compound risks: Even if a single risk source is not enough to trigger an alarm, the model can identify high-risk states under the combined action of multiple risk sources through force vector superposition, solving the limitations of traditional methods that "treat the headache and ignore the foot pain".

[0113] Implement predictive warning: The trigger of the warning is not based on "having entered" the danger zone, but on "being under strong risk gravity", achieving the forward shift of warning from event results to risk causes.

[0114] Provide precise intervention measures: The warning is no longer a simple "danger" warning, but a specific and executable "move in which direction" navigation instruction, greatly improving the effectiveness of the warning.

[0115] In order to verify the beneficial effects of the invention, the following verification report is additionally conducted: 1. Purpose of the test: The purpose of this test is to simulate real construction site environment and collect data to verify the superiority of the "a construction project site construction visual management method based on Internet of Things perception" described in the invention over the prior art. Specifically, the following three core technical effects are verified: Effect one (composite risk identification ability): verify that the method of the invention can identify and quantify the composite risk generated by the coupling of multiple risk sources, which cannot be found by existing technologies (such as electronic fence).

[0116] Effect two (early warning foresight): verify that the early warning time of the method of the invention is earlier than that of existing technologies, which can realize the early warning of "event results" to "risk causes".

[0117] Effect three (active intervention effectiveness): verify that the "navigation type risk avoidance instruction" provided by the invention can more effectively guide personnel to avoid risks compared with traditional alarms.

[0118] 2. Test design and grouping: Test site: select a building floor that has completed structural topping and is undergoing internal construction, and demarcate a 20m x 20m test area.

[0119] Test object: 1 test personnel (simulated worker) wearing an intelligent safety helmet integrated with a UWB high-precision positioning module and a heart rate sensor.

[0120] Test grouping: Control group (existing technology): deploy a traditional safety management system based on static electronic fence in the test area. The system only triggers an alarm when the real-time coordinates of the test personnel enter the preset dangerous area boundary.

[0121] Test group (method of the invention): deploy the system described in the invention in the same area, and use the CE-CRPF model for real-time risk calculation and early warning.

[0122] 3. Test scenario and parameter setting: In the test area, set up two static risk sources and introduce a dynamic risk source: Risk source A (static - hole): coordinates (15, 10), a circular floor hole with a radius of 2 meters.

[0123] Risk source B (static - unstable pile of materials): coordinates (5, 12), an unstable pile of materials with a radius of 1.5 meters.

[0124] Risk source C (dynamic - moving suspended object): simulate an area below a moving suspended object, moving in a specific trajectory.

[0125] System parameter setting: Control group (electronic fence): Electronic fence of hole A: circular area with (15, 10) as the center and a radius of 2 meters.

[0126] Electronic fence of pile B: a circular area with (5, 12) as the center and a radius of 1.5 meters.

[0127] Alarm trigger condition: the test personnel coordinates enter the interior of any fence.

[0128] Test group (CE-CRPF model of the present application): Potential point parameters:

[0129] Coupling coefficient α ij : The coupling coefficient of workers and each risk source is 1.0.

[0130] Early warning threshold: Alert threshold (Alert): 1.0; Danger threshold (Danger): 2.0; Data acquisition frequency: 10 Hz (record data once every 0.1 seconds).

[0131] 4. Test procedure and data recording: The test personnel walks in the test area according to the predetermined trajectory. The trajectory is specially designed to pass through the "saddle point area" between the two risk sources, but always does not enter the interior of any electronic fence.

[0132] Step 1: The test personnel starts from the starting point (2, 2) and walks along the predetermined trajectory.

[0133] Step 2: The trajectory will gradually approach the area between risk sources A and B.

[0134] Step 3: The trajectory is closest to the two risk sources, with a distance of 2.5 meters from A and 2.0 meters from B (both outside the fence).

[0135] Step 4: The system background records the alarm trigger time of the two groups, the position of the test personnel, the heart rate, and the total risk force modulus calculated by the test group in real time.

[0136] 5. Test data recording and analysis: 5.1 Test original data recording table: The following table selects the key time point data during the test: Table 1: Key time point data recording:

[0137] Note: At T6, the trajectory intentionally touched the electronic fence boundary of risk source A to record the alarm time of the control group.

[0138] 5.2 Effect 1 verification: composite risk identification capability: Analysis: During the time period T2 to T4, the test person walks in the area between the hole A and the pile B. As can be seen from the "person position" column, he never enters the electronic fence range of any of the risk sources. Therefore, the control group (prior art) does not issue any alarm during the entire T0-T5 phase, and the system is judged to be "safe".

[0139] However, the test group (invention method) at time T2 (15.0 seconds), the total risk force module calculated reaches 1.12, exceeding the "alert threshold" of 1.0, and triggering the "alert" level warning. This is because the CE-CRPF model correctly vectorially superimposes the "risk attraction" from the two risk sources, quantifying the composite risk of the "saddle point area" that appears to be safe but is actually accumulating risk.

[0140] Conclusion: The invention method successfully identifies the composite risk that the existing electronic fence technology cannot find, verifying its superior risk identification capability.

[0141] 5.3 Effect 2 verification: pre-warning foresight: In order to verify the pre-warning foresight, let the test person walk straight towards the risk source A at time T6.

[0142] Analysis: Control group alarm time: T6 = 30.0 seconds. The alarm occurs at the moment the person's foot steps into the electronic fence boundary. This is a typical "event result" driven lagging alarm.

[0143] Test group alarm time: T2 = 15.0 seconds. The invention method issues an "alert" warning when the person is still several meters away from the dangerous area boundary, but is already in the intersection of multiple risk source "gravitational fields".

[0144] Data comparison table 2: comparison of warning time:

[0145] Conclusion: The invention method's warning time is 15.0 seconds earlier than the existing technology. It realizes the transition from passive response to "crossing" behavior to active prediction of "towards danger" trend, and the pre-warning foresight is fully verified.

[0146] 5.4 Effect 3 verification: effectiveness of active intervention: To verify the intervention effect, two kinds of intervention are performed on the test person at time T3 (18.0 seconds), and their subsequent behavior is recorded.

[0147] Intervention A (traditional alarm): a loud alarm sound and a voice of "Danger, be careful!" in the intercom.

[0148] Intervention B (navigation instruction of the invention): a three-dimensional arrow appears on the AR glasses (or mobile phone APP) worn by the test personnel, pointing to the opposite direction of the total risk force vector, accompanied by a voice: "Please move to your left rear!" Test result Table 3: risk avoidance effect under different interventions:

[0149] Conclusion: after using the "navigation type risk avoidance instruction" of the invention, the risk avoidance reaction time of the test personnel is shortened by about 65% (from 5.2 seconds to 1.8 seconds). This proves that the active and explicit intervention instruction provided by the invention can significantly improve the efficiency and accuracy of risk avoidance compared with the traditional vague alarm, and can nip potential accidents in the bud.

[0150] 6. Test summary: Through the above three sets of comparative tests and detailed data analysis, the test verifies the excellent technical effects claimed by the technical scheme of the invention: It can effectively identify and quantify the composite risks in the blind area of the prior art.

[0151] Its early warning mechanism has significant foresight, and can gain valuable emergency disposal time compared with traditional methods.

[0152] The active intervention measures it provides are accurate and efficient, which can significantly improve the risk avoidance efficiency of personnel.

[0153] It should be noted that the above examples are only used to illustrate the technical solutions of the present application and not to limit it. Although the present application has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present application can be modified or replaced by equivalents without departing from the spirit and scope of the technical solutions of the present application, and they should be covered in the scope of the claims of the present application.

Claims

1. A method for visual management of on-site construction in a construction project using Internet of Things (IoT) sensing, characterized in that: The following steps are required: a) Construct a digital twin base model that includes a building information model, geographic information system data, and a 3D site model; b) Deploy an Internet of Things (IoT) sensing network at the construction site to acquire real-time sensing data representing the status of personnel, machinery, materials, and the environment, and establish a semantic mapping relationship between the real-time sensing data and the model units in the digital twin base model; c) Based on the mapping relationship, use the real-time sensing data to drive the digital twin base model to generate a dynamically updated digital twin in real time; d) Perform risk quantification steps, which include: d1) Abstract the dynamic construction elements in the dynamic digital twin as potential points, and assign each potential point P... i Determine its instantaneous risk potential V i (t) and real-time position vector ; d2) Based on the preset coupling type coefficient α ij Given the attenuation exponent n due to distance, calculate the potential points P for any two points P. i and P j Risk force vector between ; in: ； d3) Generate early warning instructions based on the calculated risk force vector; e) The information generated by the risk quantification step is overlaid on the visualization interface of the dynamic digital twin.

2. The IoT-based on-site construction visualization management method for construction management projects according to claim 1, characterized in that: In step d1), the instantaneous risk potential V i (t) through model V i (t)=V int,i *(1+δ i (t) is calculated; Among them, V int,i As the pre-defined inherent risk potential energy, δ i (t) is the state disturbance coefficient determined based on the real-time sensing data.

3. The IoT-based on-site construction visualization management method for construction management projects according to claim 2, characterized in that, Step d3) specifically includes: Calculate any potential point P k Total risk force ; in: ； When the magnitude of the total risk force exceeds a first preset threshold, an early warning instruction containing the potential point identifier is generated.

4. The IoT-based on-site construction visualization management method for construction management projects according to claim 3, characterized in that: The warning instruction further includes provisions based on the total risk force. The direction vector data determined in the opposite direction.

5. The IoT-based on-site construction visualization management method for construction management projects according to claim 4, characterized in that, The risk quantification step d) also includes: The composite risk potential of the predetermined set of points or regions in the construction site space is calculated based on the model. ; When the combined risk potential of any region exceeds the second preset threshold, an early warning instruction containing the location information of that region is generated. in: 。 6. The IoT-based on-site construction visualization management method for construction management projects according to claim 5, characterized in that, Step e) specifically includes: based on the composite risk potential The calculation results are used to render a three-dimensional risk heat map in the three-dimensional space of the dynamic digital twin.

7. The IoT-based on-site construction visualization management method for construction management projects according to claim 6, characterized in that, Step e) specifically includes: in response to the user's selection operation of any potential point in the dynamic digital twin, visually rendering one or more risk force vectors acting on that potential point as the center. .

8. The IoT-based on-site construction visualization management method for construction management projects according to claim 7, characterized in that, The method further includes a virtual-real deviation analysis step, which includes: Obtain a 3D point cloud model of the actual scene; The point cloud model is registered with the corresponding part of the digital twin base model and the geometric deviation is calculated. In the dynamic digital twin, the area where the deviation value exceeds the third preset threshold is highlighted.

9. The IoT-based on-site construction visualization management method for construction management projects according to claim 8, characterized in that, The method further includes a risk trend prediction step, which includes: Collect composite risk potential of a key area Values ​​within a continuous time period form time series data; The time series data is input into a time series forecasting model to output a predicted value of the future composite risk potential of the region. When the predicted value exceeds the second preset threshold, an early warning instruction is generated.

Images