BIM-based steel structure modular construction carbon footprint dynamic monitoring method

Abstract

This invention relates to a BIM-based dynamic monitoring method for the carbon footprint of modular steel structure construction, belonging to the field of carbon emission control technology in building construction. It aims to solve the problems of lagging carbon footprint accounting and the inability to dynamically control it during construction. Key aspects of this invention include: integrating progress, machinery, and transportation information into a BIM platform and pre-linking carbon factors to generate a predicted carbon footprint baseline; collecting real-time data and, based on the mapping relationship between the BIM model and physical equipment, performing data cleaning and preliminary conversion, forming a comprehensive record through spatiotemporal fusion, calculating carbon emission intensity, and comparing it with the baseline; when the cumulative deviation exceeds the limit, initiating construction scheme simulation optimization, adjusting machinery, procedures, or transportation plans, issuing new instructions after simulation verification, and reconstructing the future baseline. This invention realizes the transformation of carbon emission from static accounting to dynamic and precise control, significantly reducing carbon emissions throughout the entire process of modular steel structure construction.

Description

Technical Field

[0001] This invention relates to the field of carbon emission control technology in building construction. More specifically, this invention relates to a BIM-based method for dynamic monitoring of the carbon footprint of modular steel structure construction. Background Technology

[0002] Modular steel structure construction, as an efficient construction method, has improved construction speed and quality, but its resource consumption and carbon emissions during construction have also attracted increasing attention. Effective monitoring of the carbon footprint during the construction phase is a crucial step in achieving green and low-carbon development in the construction industry. However, the application of existing technologies in this field still faces several unresolved challenges, hindering the real-time nature and effectiveness of carbon footprint management.

[0003] First, at the planning and forecasting level, existing methods largely rely on static carbon emission accounting. Typically, before construction begins, a fixed emission coefficient per unit of work is used to estimate the total emissions based on the design model and overall schedule. This estimation usually results in a static baseline or a total value, failing to be deeply integrated with a detailed, time-evolving four-dimensional (4D) construction schedule, and also failing to fully consider the actual impact of dynamic factors such as the dynamic scheduling of construction machinery and real-time changes in the transportation paths of modular components on carbon emission intensity. Therefore, the resulting carbon footprint predictions often fail to accurately depict the true trend of carbon emission fluctuations over time throughout the entire construction process, leading to insufficient guidance in the early forecasts and inadequate comparability with the actual progress later on.

[0004] Secondly, a significant "information silo" phenomenon exists in the data collection and correlation during the construction process. Currently, collecting real-time data such as construction machinery energy consumption and vehicle GPS location via IoT sensors is technically feasible. However, these data streams typically exist independently, lacking structured association with specific components, construction tasks, and schedule nodes in the Building Information Model (BIM). Due to the lack of a unified and clear identification mapping relationship between physical equipment and BIM model objects, massive amounts of real-time operational data cannot be automatically and accurately attributed to the specific construction activity that caused the carbon emissions (such as hoisting a module with a specific number, or the continuous operation of a machine under specific conditions). This disconnect between data and model semantics prevents the real-time collected data from being effectively aligned and integrated with planned information within a unified spatiotemporal framework, creating a fundamental obstacle to subsequent dynamic analysis.

[0005] Furthermore, existing methods for real-time carbon emission calculation and process feedback suffer from significant lag and coarse-grained calculations. A common practice is to periodically (e.g., daily or weekly) aggregate various energy consumption data for subsequent batch calculations and report generation. This batch processing model cannot support real-time calculations of carbon emission intensity for specific construction activities, measured in minutes or hours. Simultaneously, due to the lack of real-time calculation and comparison capabilities, managers struggle to quickly identify deviations from expected carbon emissions while construction is underway. The invisibility of the process and the delay in feedback mean that management measures are often only reactive, failing to provide in-process early warning and immediate control.

[0006] Finally, when persistent deviations in carbon emissions are detected, dynamically optimizing and adjusting the construction plan presents significant challenges. Due to the lack of a collaborative simulation environment that integrates current progress, real-time resource status, existing carbon emission data, and carbon footprint calculation models, any adjustments (such as changing machinery combinations, adjusting the sequence of work processes, or optimizing transportation routes) struggle to quickly and quantitatively assess their combined impact on the remaining project's carbon emissions, schedule, and costs. Optimization decisions are highly experience-dependent and involve lengthy evaluation cycles, making it difficult to achieve rapid, closed-loop feedback and adjustments during construction to continuously keep carbon emissions within the expected range.

[0007] In summary, existing technologies suffer from prominent technical problems in carbon footprint management of modular steel structure construction due to static prediction models, disconnect between real-time data and models, and lagging feedback control. These problems include severely delayed calculation results and the inability to conduct dynamic monitoring and closed-loop optimization control during construction. Summary of the Invention

[0008] One object of the present invention is to solve at least the above-mentioned problems and to provide at least the advantages that will be described later.

[0009] Another objective of this invention is to provide a BIM-based dynamic monitoring method for the carbon footprint of modular steel structure construction. This method aims to solve the technical problems in existing carbon footprint management for modular steel structure construction, such as the serious lag in carbon accounting results caused by static prediction models, the disconnect between real-time data and models, and the lag in feedback control, which prevents dynamic monitoring and closed-loop optimization control during construction.

[0010] To achieve these objectives and other advantages of the present invention, a BIM-based method for dynamic monitoring of the carbon footprint of modular steel structure construction is provided, comprising: S1. Integrate the 4D schedule, machinery resource library and transportation route for modular steel structure construction into the BIM platform, pre-associate carbon footprint factors for components, machinery and transportation tasks in the model, and generate a predicted carbon footprint baseline with time as the axis through construction simulation. S2. During construction, mechanical energy consumption and transportation data are collected through IoT devices, and based on the identifier mapping relationship between the BIM model and physical equipment, real-time data streams are dynamically bound to the corresponding construction objects; streaming computation is performed based on the bound data, specifically: S21. The edge computing unit at the construction site receives the raw high-frequency data collected by the mechanical energy consumption sensor, performs data cleaning, working condition identification, and preliminary carbon equivalent conversion based on the preset carbon footprint factor, and generates a standardized unit carbon footprint data package. S22. Align and spatiotemporally fuse the unit carbon footprint data packets, transportation paths, and status data of each edge computing unit with the identification mapping relationship within a unified operation time window to form a comprehensive data record. S23. Based on comprehensive data records and pre-correlated carbon footprint factors, dynamically calculate the actual carbon emission intensity per unit time for each construction activity, and compare it in real time with the predicted intensity at the corresponding time point in the predicted carbon footprint baseline, and output the dynamic deviation. S3. When the cumulative dynamic deviation exceeds the preset threshold and reaches the preset number of times, the BIM platform automatically starts the construction scheme simulation optimization with the current state as the starting point and the remaining project as the scope; dynamically adjusts the machinery scheduling, process sequence or transportation scheme in the BIM-4D environment, and simulates and calculates the carbon footprint effect after the adjustment; issues the scheme that meets the optimization goal as a new instruction, and simultaneously reconstructs the curve of the future period in the predicted carbon footprint baseline.

[0011] Preferably, in the BIM-based dynamic monitoring method for the carbon footprint of modular steel structure construction, step S21, which involves identifying the working conditions and performing preliminary carbon equivalent conversion based on pre-set carbon footprint factors to generate standardized unit carbon footprint data packages, specifically includes: S211. The multi-source sensor group integrated on the construction machinery collects real-time energy consumption data and working condition data. Combined with the pre-set mechanical operation mode classifier, it identifies the current working condition. The working condition data includes attitude data, load data, and control signal data. S212. Call the benchmark carbon footprint factor that matches the current working condition, and use the dynamic carbon footprint factor correction model to correct the benchmark carbon footprint factor in real time to generate the dynamic carbon footprint factor; the input of the correction model includes at least the current working condition, real-time energy consumption data, and the benchmark value of energy consumption per unit time of the machine under the current working condition based on historical data. S213. Utilize the dynamic carbon footprint factor to perform carbon equivalent conversion on real-time energy consumption data and generate a unit carbon footprint data package, which includes at least the construction machinery identifier, timestamp, operating conditions, dynamic carbon footprint factor, and calculated carbon emission equivalent.

[0012] Preferably, the BIM-based dynamic monitoring method for the carbon footprint of modular steel structure construction further includes a root cause tracing step after outputting the dynamic deviation in step S23: S231. Construct and maintain a multidimensional construction status dataset, which includes at least actual carbon emission intensity, 4D progress status, resource status and utilization rate of each machine, process sequence, and environmental monitoring data. S232. When the dynamic deviation exceeds the preset threshold, based on the causal rule base and correlation analysis algorithm, the dynamic deviation is spatiotemporally correlated and pattern matched with the abnormal variables in the multidimensional construction status dataset to identify the dominant variable causing the deviation. S233. Based on the dominant variables and their influence weights, generate a root cause analysis report and output the causes of carbon footprint deviation in probability order. The types of causes of carbon footprint deviation include uneconomical operation of machinery, idle resources caused by process conflicts, congestion of transportation routes, or discrepancies between actual construction conditions and model presets.

[0013] Preferably, in the BIM-based dynamic monitoring method for the carbon footprint of modular steel structure construction, step S3, before issuing a new instruction for a scheme that meets the optimization objective, further includes: S31. Using the current BIM model status, the real-time site status collected by the Internet of Things, and the optimization scheme as inputs, construct a high-fidelity synchronous simulation environment in the digital twin layer. S32. Accelerate the pre-execution optimization scheme in a synchronous simulation environment, simultaneously simulate construction progress, resource movement and carbon emissions, and detect whether there are resource space conflicts, path conflicts or violations of safety rules during the pre-execution process. S33. If the pre-run passes and the carbon footprint optimization effect is verified, the optimization scheme will be deconstructed into a set of cooperative operation instructions with time constraints. S34. Distribute the collaborative operation instruction set to the corresponding mechanical controllers or personnel mobile terminals through the construction management platform to ensure that the instructions take effect and are executed after the set unified time reference point.

[0014] Preferably, the BIM-based dynamic monitoring method for carbon footprint of modular steel structure construction establishes the identification mapping relationship in the following way: RFID or QR code tags containing unique model identification codes are assigned to the construction machinery, transport vehicles, and prefabricated components involved in the construction; after the on-site identification equipment reads the tag information, it automatically matches it with the identification code of the corresponding object in the BIM model to form and maintain a dynamic mapping relationship table.

[0015] Preferably, in the BIM-based dynamic monitoring method for carbon footprint of modular steel structure construction, step S22 involves aligning and spatiotemporally fusing data within a unified work time window based on the identifier mapping relationship. Specifically, this involves deploying a unified reference time server at the construction site to provide a unified time stamp for unit carbon footprint data packages and transportation status data. During data alignment, the data of the same physical object within a preset time tolerance range are fused based on the time stamp to generate a comprehensive data record with a unified time reference.

[0016] Preferably, in the BIM-based dynamic monitoring method for the carbon footprint of modular steel structure construction, the method for constructing and updating the dynamic carbon footprint factor correction model in step S212 is as follows: S2121. Based on historical data, establish a machine learning model for each type of construction machinery under different operating conditions, with key operating parameters as input and unit energy consumption correction coefficient as output. Key operating parameters include at least real-time load rate, engine speed, and working pressure. S2122. Input the current operating conditions and corresponding real-time operating parameters into the machine learning model and output the unit energy consumption correction coefficient. S2123. Multiply the baseline carbon footprint factor by the unit energy consumption correction coefficient to obtain the dynamic carbon footprint factor at the current calculation time. S2124. Regularly use newly added actual energy consumption and operating condition data as training samples to iteratively update the machine learning model.

[0017] Preferably, in the BIM-based dynamic monitoring method for the carbon footprint of modular steel structure construction, the optimization target in step S3 is used to issue new instructions through the following hierarchical mechanism: a) Decompose the optimization scheme into automated execution instructions and manually assisted execution instructions; b) For construction machinery connected to a controller and capable of automatic execution, the automatic execution command will be directly sent to its controller; c) For processes requiring manual intervention, push the manual assistance instructions to the mobile terminals of relevant personnel; d) All instructions are accompanied by a unified effective time and are simultaneously visualized in the BIM platform.

[0018] The present invention has at least the following beneficial effects: 1. This invention constructs a complete closed-loop management and control chain, from pre-generation of carbon footprint baselines to real-time monitoring and scheme optimization, solving the problems of lagging traditional carbon accounting and inability to dynamically control emissions. Specifically: by integrating 4D progress, machinery, and transportation information and associating them with carbon factors through a BIM platform, the generated timeline baseline can accurately anchor the expected carbon emissions for each period; spatiotemporal fusion and dynamic deviation comparison of real-time data enable real-time monitoring of carbon emissions during construction; simulation optimization and baseline reconstruction after deviations exceed limits can quickly bring carbon emissions back to the expected trajectory, promoting the transformation of carbon management in modular steel structure construction from static accounting to dynamic and precise governance, and improving the controllability of carbon emission reduction throughout the entire process.

[0019] 2. This invention refines the generation logic of unit carbon footprint data packages, solving the drawback of traditional fixed carbon factors being unable to adapt to dynamic mechanical operating conditions. Specifically: by identifying different operating conditions such as no-load idling and hoisting operations through multi-source sensors, and combining with a dynamic carbon footprint factor correction model, the carbon factor can be adjusted according to parameters such as real-time energy consumption and operating condition baseline values, making the carbon equivalent conversion more consistent with the actual operating state of the machinery; the standardized data package contains complete key information such as machinery identification and operating conditions, providing a high-quality basic unit for subsequent multi-source data fusion and accurate carbon emission calculation, significantly improving the accuracy and standardization of front-end data collection.

[0020] 3. This invention enhances the ability to trace the root causes of carbon footprint deviations, solving the problem of accurately pinpointing the causes after deviations occur. Specifically, by constructing a multi-dimensional construction status dataset, integrating information such as carbon emission intensity, progress status, and machinery utilization, and relying on a causal rule base and correlation analysis algorithm, it can quickly match the spatiotemporal correlation between deviations and abnormal variables, identifying dominant causes such as uneconomical machinery operation and process conflicts. The generated root cause report presents problems in probability order, providing managers with clear directions for rectification, avoiding blind adjustments, and improving the targeting and efficiency of carbon footprint deviation management.

[0021] 4. This invention adds a digital twin pre-simulation step before the optimized plan is issued, solving the problems of conflicts and safety hazards that easily occur during the implementation of the optimized plan. Specifically: Based on the BIM model and the real-time site status, a high-fidelity simulation environment can be built to predict the progress, resource movement and carbon emissions after the plan is implemented, and accurately detect spatial and path conflicts and safety violations; through instruction set deconstruction and precise distribution, it ensures that the plan can take effect according to a unified time benchmark, which not only verifies the carbon footprint optimization effect, but also ensures construction safety and collaboration, and improves the reliability and safety of the optimized plan implementation.

[0022] 5. This invention improves the construction and iteration mechanism of the dynamic carbon footprint factor correction model, solving the problems of insufficient adaptability and difficulty in continuous optimization of carbon factors. Specifically: based on historical data, machine learning models are established for various types of machinery under different operating conditions. Energy consumption correction coefficients can be output according to key parameters such as load rate and engine speed, making the dynamic carbon factor more consistent with the real-time operating status of the machinery; the model is iterated regularly with new data, which can continuously improve the prediction accuracy of the model, avoid the limitations of fixed carbon factors, and ensure the long-term accuracy of carbon equivalent conversion from the underlying algorithm level.

[0023] 6. This invention establishes a hierarchical instruction issuance mechanism, solving the problems of poor coordination in the execution of optimization schemes and low efficiency in instruction implementation. Specifically: by decomposing the scheme into two types of instructions—automated and manual—it can adapt to the execution needs of intelligent machinery and manual processes respectively; automated instructions are directly connected to the machine controller, while manual instructions are pushed to mobile terminals, and all instructions come with a unified effective time, ensuring the synchronization of construction rhythm at each stage; the visualization display of the BIM platform enables real-time traceability of instruction execution status, significantly improving the coordination and controllability of the implementation of optimization schemes.

[0024] Other advantages, objectives and features of the present invention will become apparent in part from the following description, and in part from those skilled in the art through study and practice of the invention. Detailed Implementation

[0025] The present invention will be further described in detail below with reference to embodiments, so that those skilled in the art can implement it based on the description.

[0026] It should be understood that terms such as “having,” “comprising,” and “including” as used herein do not exclude the presence or addition of one or more other elements or combinations thereof.

[0027] It should be noted that, unless otherwise specified, the experimental methods described in the following implementation plan are all conventional methods, and the reagents and materials described are all commercially available unless otherwise specified.

[0028] In one embodiment of the present invention, a method for dynamic monitoring of the carbon footprint of modular steel structure construction based on BIM is provided, comprising: S1. Integrate the 4D schedule, machinery resource library and transportation route for modular steel structure construction into the BIM platform, pre-associate carbon footprint factors for components, machinery and transportation tasks in the model, and generate a predicted carbon footprint baseline with time as the axis through construction simulation. S2. During construction, mechanical energy consumption and transportation data are collected through IoT devices, and based on the identifier mapping relationship between the BIM model and physical equipment, real-time data streams are dynamically bound to the corresponding construction objects; streaming computation is performed based on the bound data, specifically: S21. The edge computing unit at the construction site receives the raw high-frequency data collected by the mechanical energy consumption sensor, performs data cleaning, working condition identification, and preliminary carbon equivalent conversion based on the preset carbon footprint factor, and generates a standardized unit carbon footprint data package. S22. Align and spatiotemporally fuse the unit carbon footprint data packets, transportation paths, and status data of each edge computing unit with the identification mapping relationship within a unified operation time window to form a comprehensive data record. S23. Based on comprehensive data records and pre-correlated carbon footprint factors, dynamically calculate the actual carbon emission intensity per unit time for each construction activity, and compare it in real time with the predicted intensity at the corresponding time point in the predicted carbon footprint baseline, and output the dynamic deviation. S3. When the cumulative dynamic deviation exceeds the preset threshold and reaches the preset number of times, the BIM platform automatically starts the construction scheme simulation optimization with the current state as the starting point and the remaining project as the scope; dynamically adjusts the machinery scheduling, process sequence or transportation scheme in the BIM-4D environment, and simulates and calculates the carbon footprint effect after the adjustment; issues the scheme that meets the optimization goal as a new instruction, and simultaneously reconstructs the curve of the future period in the predicted carbon footprint baseline.

[0029] The core of this implementation method is to integrate a 4D schedule, machinery resource library, and transportation routes for modular steel structure construction into the BIM platform, and pre-associate carbon footprint factors for various objects in the model. The 4D schedule can import construction timelines generated by project management software, with half-days or hours as the smallest unit. The machinery resource library can be established as a database containing parameters of equipment such as truck cranes, crawler cranes, and self-propelled modular transport vehicles. Transportation routes can be drawn in the BIM environment based on the site plan. Carbon footprint factors can be selected from the corresponding energy and material emission coefficients in the "Unified Standard for Calculating Building Carbon Emissions," for example, electricity consumption is based on the power grid emission factor of the province where the project is located, and diesel consumption uses a fixed carbon dioxide emission coefficient. Through the construction simulation function of the BIM software, the model objects are driven to move according to the timeline, and the system automatically calculates and generates a baseline curve for predicted carbon footprint intensity in daily or hourly units. During implementation, the BIM software and database can be installed on the project department's local server or cloud server for access and operation by management personnel.

[0030] During construction, real-time data is collected via IoT devices. For mechanical energy consumption, fuel flow meters with pulse output function can be installed on diesel-powered truck cranes, and smart meters can be installed on electric hydraulic pump stations. For transportation data, vehicle-mounted terminals integrating GPS and 4G communication functions can be installed on modular transport vehicles. To achieve dynamic binding between physical equipment and the BIM model, each major construction machine and transport vehicle can be assigned a unique digital code, which is then written into the equipment identifier of its vehicle-mounted terminal; each precast steel structure module is labeled with a QR code containing the component number. Industrial-grade wireless networks (such as Wi-Fi 6 or 5G private networks) are deployed in key work areas on site to ensure data transmission. When machinery starts or components arrive on site, their identification information and real-time data stream are synchronously transmitted to the BIM platform. The platform automatically associates and matches the data stream with the corresponding virtual objects in the BIM model according to predefined coding rules, forming a dynamically updated list of binding relationships.

[0031] Based on this binding relationship, the streaming computing process is as follows: Edge computing gateways with certain computing capabilities are deployed in the main work areas of the construction site. Sensors on the machinery send raw energy consumption data (such as flow pulses and current values) to the nearest edge gateway in real time. The gateway first verifies the data, for example, filtering out unreasonable values ​​exceeding the equipment's rated range. Then, combining auxiliary signals read from the mechanical controller bus (such as the CAN bus: Controller Area Network bus) (such as hydraulic system pressure and engine load percentage), and using simple rules pre-set within the gateway (e.g., if engine speed > 700 rpm and load sensor reading is 0, it is judged as "no-load driving"), the current main operating condition is identified. The system calls the baseline carbon footprint factor corresponding to this operating condition. To be more realistic, a linear correction method can be used: based on the real-time monitored load rate (such as the current lifting weight as a percentage of the rated lifting capacity), the corresponding correction coefficient is looked up from a pre-stored empirical coefficient table (e.g., the coefficient is 0.95 when the load rate is 50%), and the baseline factor is multiplied by this coefficient to obtain the dynamic factor used for real-time calculation. Then, this dynamic factor is multiplied by the cumulative energy consumption over that time period to calculate the preliminary carbon emission equivalent, which is then packaged into a standardized data unit containing the device ID, start and end times of the time period, operating conditions, and carbon emissions. After processing and data packet encapsulation, the edge gateway uploads it to the central server.

[0032] The central server receives data packets from each edge gateway and GPS track point data reported by transport vehicles. To ensure consistency in all data times, a unified clock source is configured on-site, and all edge gateways and data acquisition terminals synchronize their times via the network, attaching precise timestamps to each data entry. The server-side fusion program, based on the device ID and timestamp, splices and integrates data segments from the same device that are consecutive or overlapping in time. For example, it merges multiple fuel consumption data packets generated by a transport vehicle between 10:00 and 10:30 into a single complete transport operation record for that period.

[0033] Based on these fused records, the system dynamically summarizes the total actual carbon emissions generated by all construction activities within a one-hour statistical period, and calculates the average carbon emission intensity within that period. Simultaneously, the system extracts the predicted intensity value for the same hour from the prediction baseline curve generated in step S1. The actual value is compared with the predicted value to calculate the deviation (e.g., (actual value - predicted value) / predicted value * 100%). The system sets a deviation threshold, such as 20%. When the actual carbon emission intensity exceeds the predicted value and reaches this threshold, and continues to accumulate beyond a preset number of times (e.g., exceeding the limit in two consecutive statistical periods (2 hours), it is determined that optimization needs to be initiated.

[0034] Once the optimization conditions are triggered, the BIM platform automatically starts from the current model state (reflecting completed works and ongoing tasks) and performs simulation optimization on all remaining processes. In the BIM 4D simulation interface, variables can be manually adjusted or through simple algorithms, such as temporarily swapping the working areas of two mobile cranes or replanning a detour route for a transport vehicle. After each adjustment, the system uses built-in carbon emission calculation rules to quickly simulate the construction process from the current moment to the project's end, and outputs the predicted total carbon emissions and project duration changes under the new scheme. The constraints to be met for optimization include: processes on the critical path must not be delayed; the daily operating time of any machinery must not exceed the upper limit specified in its safety operating procedures.

[0035] After the simulation is completed, the system will adopt a solution that reduces subsequent carbon emission predictions without violating constraints as the new instruction. Before issuing the instruction, the solution is broken down: for intelligent transport vehicles that can receive digital instructions, a new sequence of navigation coordinate points is generated and issued; for hoisting machinery requiring manual operation, an updated list of hoisting component sequences and time suggestions are generated and pushed to the operator's handheld smart terminal (such as an explosion-proof tablet). All instructions are set with a unified future effective time, such as "3 PM today". At the same time, the predicted carbon footprint baseline in the BIM platform will be updated from this effective time according to the simulation results of the new solution, forming a new target curve for future monitoring and comparison.

[0036] This implementation establishes a closed-loop control chain of "prediction-real-time monitoring-deviation analysis-dynamic optimization-benchmark reconstruction." Existing technologies typically stop at static carbon estimation before construction or periodic carbon accounting after construction, failing to accurately and automatically aggregate the massive, heterogeneous real-time operational data generated during construction into the corresponding construction activities in the BIM model through unified identification. This hinders high-time-granularity real-time calculation and comparison of carbon emission intensity. More importantly, existing technologies lack a BIM simulation optimization environment that, after detecting excessive carbon emissions, integrates multi-dimensional constraints and can quickly quantify and assess the impact of adjustment plans on subsequent carbon footprints. Therefore, proactive, closed-loop control during construction is difficult to achieve. This solution, by connecting the entire process of data binding, streaming computation, and simulation optimization, shifts carbon footprint management during the construction phase from post-construction statistics to in-process dynamic control, improving the timeliness and effectiveness of management.

[0037] A prefabricated steel structure modular office building construction project was selected as the application object to verify the effectiveness of the BIM-based dynamic monitoring method for carbon footprint of steel structure modular construction in solving the problems of lagging traditional carbon accounting and inability to dynamically control carbon footprint.

[0038] First, a BIM-based baseline for predicting carbon footprint is constructed. A 4D project schedule is integrated into the BIM platform, using hours as the smallest time unit, specifying the exact time windows for each steel component hoisting and transportation. A machinery resource library is imported, including parameters such as the rated power and standard operating cycle time of the QTZ80 tower crane, the fuel consumption rate of 10t transport vehicles, and the mileage and typical travel time for each transportation route. Carbon footprint factors are pre-associated for each object in the BIM model: the tower crane uses a unit power consumption carbon emission coefficient based on the average emission factor of the power grid area, the transport vehicle uses a diesel emission coefficient, and component transportation is associated with unit ton-kilometer carbon emission parameters. Through BIM construction simulation, the 4D schedule is driven, and combined with machinery efficiency and route information, a baseline curve for predicting carbon footprint intensity is generated in hours. This curve reflects the expected fluctuations in carbon emissions during different construction phases each day (e.g., concentrated hoisting in the morning and dispersed welding in the afternoon).

[0039] Real-time monitoring was initiated during construction. Power consumption sensors, load sensors, and slewing angle sensors were installed on the tower crane, and GPS and fuel flow meters were installed on the transport vehicle. Data was collected every 5 minutes. A mapping between physical and digital objects was established using unique identifiers: each steel component was affixed with an RFID tag, and the tower crane and transport vehicle were associated with the BIM model through a unique ID built into their onboard intelligent terminals. Edge computing nodes deployed on-site received raw data, cleaned and verified it, and identified the mechanical operating conditions in real time based on sensor data streams (such as load values, boom angles, and vehicle speeds). Tower cranes were categorized as "idle standby," "loaded slewing," and "lifting / lowering," while transport vehicles were categorized as "driving," "idling," and "loading / unloading." After identification, the system retrieved the baseline carbon footprint factor matching the operating condition and ran a dynamic correction model in real time. This model was trained based on a large amount of historical data, inputting parameters such as real-time load rate and engine speed, and outputting an energy consumption correction coefficient for the operating condition. Multiplying the baseline factor by this coefficient yielded the dynamic carbon footprint factor. Subsequently, the carbon emission equivalent is converted in real time at the edge according to the formula "carbon emission equivalent = measured energy consumption × dynamic carbon footprint factor", generating a standardized data package with fuselage ID, timestamp, operating condition, dynamic factor and carbon emission equivalent.

[0040] All data packets generated by edge nodes and GPS trajectory data of transport vehicles are synchronized to a base time server deployed on-site, and stamped with a unified millisecond-level time stamp. On the central processing platform, based on the identifier mapping relationship, data from the same object within the same time period are spatiotemporally aligned and fused to form a comprehensive data record covering the entire site and with continuous time. Based on this, the system dynamically calculates the actual carbon emission intensity per hour and compares it in real time with the predicted value for the corresponding hour in the prediction baseline curve, generating a dynamic deviation.

[0041] When the system detects that the carbon emission intensity of a construction area exceeds the benchmark value by 15% for two consecutive hours, it automatically triggers the root cause analysis process. The system retrieves a multidimensional state dataset for that period, including machinery operating condition sequences, process progress, vehicle location and speed, and on-site video logs, and performs pattern matching using an association rule algorithm. The analysis reveals that during the period of exceeding the limit, the transport vehicle numbered Truck-02 repeatedly exhibited a state of "speed <5km / h and continuous idling" on a certain section of Route 2, and this section was the only passable route; at the same time, the tower cranes waiting for the vehicle's components showed an abnormally high proportion of "idle standby" conditions. Based on this, the system generates a root cause report, determining that the main cause is the local congestion on Route 2, which led to a decrease in transportation efficiency and resulted in the subsequent idleness of hoisting resources.

[0042] Based on this diagnosis, when the same problem causes the daily cumulative deviation to exceed a preset threshold, the BIM platform automatically initiates simulation optimization. Starting from the current actual state of the project, the remaining work processes are simulated in the BIM-4D environment: First, the transportation route is switched from route 2 to route 1 with better traffic capacity; second, the affected tower crane operation sequence is fine-tuned, and the hoisting of some non-critical path components is postponed to match the new transportation rhythm. The simulation model takes "minimizing the increase in total carbon emissions" as the main objective, and "no delay in the construction period of key nodes" and "the number of tower crane operation cycles per day does not exceed the safe limit" as constraints, and performs rapid iterative calculations. The simulation results show that the new solution can reduce the average daily carbon emissions of subsequent construction by about 8%, while keeping the total construction period unchanged.

[0043] After the optimized plan was successfully simulated in a digital twin environment, it was broken down into specific instructions: a new navigation route was issued to the Truck-02 vehicle terminal; and updated lifting sequences and time windows were pushed to the handheld terminals of relevant tower crane operators. All instructions included a unified effective time (e.g., 6:00 AM the following day). Simultaneously, the predicted carbon footprint baseline curve in the BIM platform was dynamically reconstructed from the effective time to reflect the expected carbon emission trajectory under the new plan.

[0044] The comparative method used traditional static accounting, estimating the total carbon emissions of the project solely based on the total workload and fixed emission coefficients before construction. During construction, fuel and electricity consumption data for machinery were manually collected every weekend for post-construction accounting. After discovering that emissions in the second week exceeded the weekly average budget by 22%, the decision to "increase the number of transport vehicles" was made based solely on experience, without assessing the impact on on-site traffic flow. Following implementation, the increased number of vehicles on-site led to queues, increasing tower crane waiting time and generating additional idling energy consumption. Furthermore, subsequent calculations continued to use the original total emission targets, making process control impossible.

[0045] The above verification tests show that the present invention achieves dynamic and precise control of construction carbon footprint through high time granularity baseline, real-time data fusion based on unified time scale, deviation root cause diagnosis through correlation analysis, and closed-loop optimization through simulation verification, effectively solving the problems of lagging, extensive, and blind regulation in traditional methods.

[0046] In another embodiment of the present invention, the BIM-based method for dynamic monitoring of carbon footprint in modular steel structure construction, specifically step S21, involves identifying the working conditions and performing preliminary carbon equivalent conversion based on pre-set carbon footprint factors to generate standardized unit carbon footprint data packages. S211. The multi-source sensor group integrated on the construction machinery collects real-time energy consumption data and operating condition data. Combined with a pre-set machinery operation mode classifier, it identifies the current operating condition. The operating condition data includes attitude data, load data, and control signal data. The operating condition is identified according to the type of machinery: for hoisting machinery, the operating conditions include: no-load idling, hoisting operation, and no-load movement; for transport vehicles, the operating conditions include: no-load driving, loaded driving, and loading / unloading idling. S212. Call the benchmark carbon footprint factor that matches the current working condition, and use the dynamic carbon footprint factor correction model to correct the benchmark carbon footprint factor in real time to generate the dynamic carbon footprint factor; the input of the correction model includes at least the current working condition, real-time energy consumption data, and the benchmark value of energy consumption per unit time of the machine under the current working condition based on historical data. S213. Utilize the dynamic carbon footprint factor to perform carbon equivalent conversion on real-time energy consumption data and generate a unit carbon footprint data package, which includes at least the construction machinery identifier, timestamp, operating conditions, dynamic carbon footprint factor, and calculated carbon emission equivalent.

[0047] In this embodiment, the multi-source sensor group integrated on the construction machinery is selected and installed in the following locations: For a truck crane, a turbine flow sensor with analog or pulse signal output can be installed in its fuel line to collect real-time fuel consumption data. This sensor can be installed on the pipeline after the engine fuel pump and before the high-pressure fuel pump. To obtain operating condition data, a tension sensor can be installed at the fixed end of the lifting wire rope of the crane hook to measure the lifting load; a triaxial acceleration sensor can be installed on the crane chassis frame to monitor the vehicle's posture when stationary or moving; existing control signal data, such as engine speed, hydraulic system pressure, and boom angle, can also be read from the crane's controller area network bus. These sensors can be connected to an edge computing unit fixed in the crane's cab via shielded cables or wireless transmission modules. This unit can be an industrial-grade embedded computer responsible for receiving and processing the raw data from all sensors.

[0048] The process of operating condition identification and preliminary carbon equivalent conversion based on preset carbon footprint factors is as follows: The operating condition identification program built into the edge computing unit loads the corresponding rule base according to the device type and performs real-time analysis on the collected data stream.

[0049] For lifting machinery (such as cranes), the judgment is mainly based on the tension sensor, boom angle sensor, and engine data. For example, the program can set rules: when the engine speed is higher than the idle speed threshold and the hook load is continuously lower than 5% of the rated value, it is judged as "no-load idling"; when the hook load increases significantly and is accompanied by a change in boom angle, and the vehicle posture is stable (no movement), it is judged as "lifting operation"; when the hook is unloaded and the vehicle posture sensor shows continuous movement, it is judged as "no-load movement".

[0050] For transport vehicles (such as modular transport vehicles), the judgment is mainly based on vehicle attitude sensors, load sensors, and positioning data. For example, the program can set rules: when the vehicle is moving and the load is less than 10% of the rated value, it is judged as "driving without load"; when the vehicle is moving and the load is more than 10% of the rated value, it is judged as "driving with load"; when the vehicle is stationary and the engine is running, it is judged as "idling while loading and unloading".

[0051] The program then calls a baseline carbon footprint factor that matches the current equipment type and the identified operating status. This factor value can be pre-stored in the local storage of the edge computing unit. For example, under the "idle" condition, the baseline carbon emission factor for this type of crane consuming 1 liter of diesel is a fixed value multiplied by kilograms of CO2 equivalent. To generate a dynamic carbon footprint factor, the system runs a dynamic correction model. This model can be a simple linear correction function whose inputs are the current operating condition, the real-time fuel consumption rate (liters / hour), and the baseline average fuel consumption rate of the machinery under this condition, calculated from historical data (e.g., the historical average for "lifting operations" is 15 liters / hour). The correction function can be designed as: Dynamic Factor = Baseline Factor × (Real-time Fuel Consumption Rate / Historical Average Rate). Through this calculation, if the current actual fuel consumption is higher than the historical average, the dynamic factor is increased; otherwise, it is decreased. Finally, this dynamic carbon footprint factor is used to convert real-time energy consumption data into carbon equivalents. For example, if 0.8 liters of fuel were consumed in the past 5-minute data collection period, and the dynamic factor is Y kg CO2 equivalent / liter, then the carbon emission equivalent is 0.8 × Y kg. The edge computing unit encapsulates the results of this calculation into a standardized unit carbon footprint data package.

[0052] The generated unit carbon footprint data package includes at least the construction machinery identifier (e.g., equipment number "Crane-001"), timestamp (format "YYYY-MM-DD HH:MM:SS"), operating condition (e.g., code "LIFTING"), the value of the dynamic carbon footprint factor used, and the calculated carbon emission equivalent. This data package is periodically uploaded to the central BIM platform server via a wireless network (e.g., 4G or Wi-Fi) at the construction site (e.g., every 5 minutes). The data package is serialized using lightweight JSON or Protocol Buffers format to reduce network transmission burden. On the central server, a corresponding parsing program receives and verifies these data packages, ensuring their structural integrity and data rationality, providing standardized input for subsequent spatiotemporal fusion steps.

[0053] This implementation method shifts the calculation of machinery carbon emissions from using a single fixed factor to a refined management model that combines real-time multi-sensor data to identify specific operating conditions and dynamically adjusts the carbon factor based on current energy efficiency. By integrating multi-source sensors locally on the machinery and performing edge computing, this implementation method can more accurately depict the actual carbon emission characteristics of machinery under different operating conditions. This makes the carbon equivalent calculated for each machine and each operating period closer to reality, thus providing a more reliable basic data unit for accurate and dynamic monitoring of project-level carbon footprint.

[0054] In another embodiment of the present invention, the BIM-based dynamic monitoring method for carbon footprint of modular steel structure construction further includes a root cause tracing step after outputting the dynamic deviation in step S23: S231. Construct and maintain a multidimensional construction status dataset, which includes at least actual carbon emission intensity, 4D progress status, resource status and utilization rate of each machine, process sequence, and environmental monitoring data. S232. When the dynamic deviation exceeds the preset threshold, based on the causal rule base and correlation analysis algorithm, the dynamic deviation is spatiotemporally correlated and pattern matched with the abnormal variables in the multidimensional construction status dataset to identify the dominant variable causing the deviation. S233. Based on the dominant variables and their influence weights, generate a root cause analysis report and output the causes of carbon footprint deviation in probability order. The types of causes of carbon footprint deviation include uneconomical operation of machinery, idle resources caused by process conflicts, congestion of transportation routes, or discrepancies between actual construction conditions and model presets.

[0055] This implementation provides a root cause tracing step for deviations. This step is initiated after the system outputs dynamic deviations. The aim is to build a data environment that can integrate the multi-dimensional status of construction and, based on this, locate the root cause of carbon emission deviations through analysis.

[0056] First, the multidimensional construction status dataset is constructed and maintained. This dataset is created and updated in a dedicated database on a central server or cloud platform. The dataset integrates data streams from different systems: actual carbon emission intensity data comes from the real-time calculation results of step S23; 4D progress status data is obtained from a BIM platform integrated with the schedule plan, which may include comparison markers of the planned start and completion times and actual status of each construction task (such as "lifting of the third-floor steel beam in Area A"); resource status and utilization rate data of each machine comes from unit carbon footprint data packets reported by the IoT edge computing unit and machine controller status signals, for example, a tower crane operating in "lifting operation" mode for 70% of the time and "idling" mode for 30% of the time between 8:00 and 10:00 AM; process sequence data is synchronized from the project construction management software; environmental monitoring data is provided by an independent sensor network deployed on the construction site. These sensors may include online dust monitors (monitoring PM2.5 and PM10 data) installed around the construction site and ultrasonic anemometers installed on the top of the tower crane or at the highest point of the construction site. All data is timestamped through a unified time server and associated with a unified construction object identifier (such as component number, machinery number) to form a multi-dimensional status snapshot table covering "progress-resources-environment", which is then updated on a rolling basis at a set time frequency (such as every 15 minutes).

[0057] When the dynamic deviation exceeds a preset threshold, the system automatically performs root cause analysis. This preset threshold can be set as follows: the actual carbon emission intensity exceeds the predicted baseline value by 20% for two consecutive statistical periods (e.g., two consecutive hours). The system retrieves snapshots of the multidimensional construction status dataset for the period when the deviation occurred (e.g., 14:00-16:00 the previous day) and within the preceding and following time windows. Subsequently, the system calls a pre-set causal rule base and an association analysis algorithm for pattern matching. The causal rule base contains a series of empirical rules in the form of "IF-THEN," such as: "IF average vehicle speed on a certain transportation route < 5 km / h AND 'the associated hoisting machinery vacancy rate on this route > 40%, THEN root cause probability: transportation congestion leads to resource idleness' increases." The association analysis algorithm (such as the Apriori algorithm) is used to mine the spatiotemporal association rules between various abnormal variables in the dataset (e.g., "abnormally high vacancy rate of machinery A," "sudden increase in vehicle density on route B," "PM10 concentration exceeding the standard in a specific area") and the carbon emission intensity deviation. By comparing and weighting the dynamic deviations with these outlier variables, the system can identify one or more dominant variables that cause the deviations.

[0058] Based on the identified dominant variables and their influence weights calculated by the algorithm, the system automatically generates a structured root cause analysis report. The report outputs the most likely causes of carbon footprint deviation, sorted by probability from highest to lowest. For example, the report might output: "1. Root Cause: Uneconomical Operation of Machinery (Probability: 65%). Details: During the deviation period, the Crane-102 tower crane operated at 'no load idling' for 45% of the time, far exceeding the historical average of 15%, suggesting excessively long waiting times for components. Related Data: GPS data shows that the vehicle transporting the components scheduled to be lifted by this tower crane is stuck on the road outside the plant." or "2. Root Cause: Transportation Route Congestion (Probability: 70%). Details: During the deviation period, the average travel time for the 'South Gate to Stockyard' route increased from 8 minutes to 22 minutes, causing disruption to the subsequent lifting operation chain. Related Data: During this period, the number of vehicles traveling on this route increased from the normal 3 to 8, and environmental monitoring showed no abnormal weather in the area." The report is pushed to project managers, dispatchers, and other relevant management personnel through the construction management platform's visual interface, providing them with clear directions for investigation and intervention.

[0059] This implementation adds automated, data-driven root cause diagnosis capabilities to the construction carbon emission monitoring system. Existing technologies, such as some smart construction site platforms, while enabling online monitoring and alarms for exceeding carbon emission limits, still require managers to rely on personal experience and manually review isolated logs from different systems (such as progress reports, vehicle GPS records, and machinery operation records) to guess the cause of the deviation after an alarm occurs. This process is time-consuming, subjective, and prone to missing key correlations. This solution, by pre-constructing a dataset integrating multi-dimensional information and applying rules and algorithms for automatic correlation analysis, directly links "deviation phenomena" with "abnormal construction status," quickly locating the source of the problem, such as improper use of machinery, malfunctioning work processes, or external traffic issues. This changes the inefficient traditional post-event manual investigation model, shifting carbon emission process control from simple "monitoring and alarming" to "diagnostic attribution," significantly improving the speed of response and the targeted nature of intervention measures for managers regarding carbon footprint deviations.

[0060] In one embodiment of the present invention, the dynamic monitoring method for carbon footprint of modular steel structure construction based on BIM, in step S3, before issuing a new instruction for a scheme that meets the optimization objective, further includes: S31. Using the current BIM model status, the real-time site status collected by the Internet of Things, and the optimization scheme as inputs, construct a high-fidelity synchronous simulation environment in the digital twin layer; the digital twin layer refers to a virtual simulation environment driven by the current BIM model status and real-time Internet of Things data, used for high-precision simulation and forward-looking verification. S32. Accelerate the pre-execution optimization scheme in a synchronous simulation environment, simultaneously simulate construction progress, resource movement and carbon emissions, and detect whether there are resource space conflicts, path conflicts or violations of safety rules during the pre-execution process. S33. If the pre-run passes and the carbon footprint optimization effect is verified, the optimization scheme will be deconstructed into a set of cooperative operation instructions with time constraints. S34. Distribute the collaborative operation instruction set to the corresponding mechanical controllers or personnel mobile terminals through the construction management platform to ensure that the instructions take effect and are executed after the set unified time reference point.

[0061] In this embodiment, the construction of the digital twin layer synchronous simulation environment can be achieved based on the secondary development interface between a commercial BIM software platform and a game engine. Once the optimization scheme is generated, the system uses the status of completed and ongoing components in the current BIM model, real-time on-site mechanical coordinates and working condition data collected by the Internet of Things, and the optimization scheme including adjustment measures (such as a modified transportation route sequence) as input to drive the initialization of the simulation environment. For example, in the simulation environment, the virtual transport vehicle will start from the GPS location last reported by its physical vehicle and travel along the newly planned route; the virtual tower crane will start from its currently recorded boom angle and height and execute a new lifting task sequence. This simulation environment runs on a dedicated graphics workstation or cloud server within the project department.

[0062] In a synchronous simulation environment, the system pre-runs the optimization plan at an acceleration factor (e.g., 10x or 50x compared to actual time). The simulation engine synchronously extrapolates the virtual construction progress and resource movement trajectories, and calls the integrated carbon emission calculation model to estimate the carbon footprint changes during the pre-run period. Simultaneously, a collision detection algorithm continuously runs, detecting resource space conflicts (e.g., the booms of two virtual tower cranes entering each other's safe radius at a certain moment), path conflicts (e.g., virtual transport vehicles intersecting in narrow sections), or violations of safety rules (e.g., virtual components getting too close to high-voltage power lines on the hoisting path). These detection results are logged in real time; if a conflict is detected, the pre-run is interrupted and the plan is flagged as risky.

[0063] If the simulation passes and the carbon footprint optimization effect meets expectations (e.g., the cumulative carbon emission estimate at the end of the day is lower than the original plan), the system will deconstruct the optimization scheme into a set of collaborative operation instructions with time constraints. For example, one instruction is "Truck-05, execute at 14:00, navigate to path point sequence [P1, P2, P3]"; another instruction is "Tower crane Operator-03, execute at 14:15, hoist component B-207". Subsequently, the construction management platform will directly send automated execution instructions (such as navigation paths) to the on-board controllers of the corresponding transport vehicles via 4G / 5G networks or the construction site LAN, and push manual execution instructions to the industrial-grade mobile terminals (such as ruggedized tablets) of the relevant operators. All instructions are accompanied by a unified future effective time reference point (e.g., "14:00 of the day") and are simultaneously highlighted and visualized in the BIM platform view, facilitating global monitoring of the instruction status by management personnel.

[0064] This implementation adds a high-fidelity pre-simulation and conflict detection step based on digital twins before the optimized solution is issued. In existing technologies, even after an optimized solution is generated, it is usually directly issued for execution or only simple logical checks are performed. There is a lack of a simulation environment that can integrate the current precise on-site conditions and simulate the entire execution process of the solution in virtual space in advance. This results in potential spatial conflicts, path conflicts, or safety hazards in the solution not being detected in advance, which may lead to construction interruptions and safety accidents during actual execution, causing additional resource waste and carbon emissions. This implementation, by constructing a synchronous simulation environment and conducting accelerated pre-simulation, can verify the feasibility and safety of the solution before implementation, ensuring that the carbon footprint optimization solution itself is robust and executable, thereby improving the success rate of closed-loop control and on-site safety.

[0065] In another embodiment of the present invention, the BIM-based dynamic monitoring method for carbon footprint of modular steel structure construction establishes the identification mapping relationship in the following way: RFID or QR code tags containing unique model identification codes are assigned to the construction machinery, transport vehicles and prefabricated components involved in the construction; after the on-site identification equipment reads the tag information, it automatically matches it with the identification code of the corresponding object in the BIM model to form and maintain a dynamic mapping relationship table.

[0066] In this embodiment, to achieve identification mapping between physical equipment and BIM model objects, labels containing unique model identifiers can be assigned to construction machinery, transport vehicles, and prefabricated components. Construction machinery, such as crawler cranes, can be equipped with passive UHF RFID tags, which can be affixed to the inside of the cab doors and windows or near the equipment nameplate. Transport vehicles, such as heavy modular transport vehicles, can have QR code labels with unique numbers affixed to the inside of their windshields. For prefabricated steel structure components, a weather-resistant RFID tag can be welded or bound to a location on the side that is unlikely to be obscured by subsequent construction before they leave the factory, or a QR code containing the component number can be sprayed onto the component surface. The information stored or encoded in these tags is completely consistent with the "Asset ID" or "Component Number" field of the corresponding object in the BIM model.

[0067] On-site identification equipment includes fixed RFID readers and handheld smart terminals. Fixed RFID readers can be installed at key locations such as main entrances and exits, material yard entrances, and tower crane operator cabins on the construction site. The reading distance can be selected from models with 3-5 meters or 5-10 meters, depending on the needs. Construction personnel or managers can be equipped with industrial-grade handheld terminals or explosion-proof smartphones with RFID reading and QR code scanning capabilities. When tagged machinery, vehicles, or components enter the reader's identification area, or are scanned by a handheld terminal, the identification device reads the unique identifier from the tag and sends the "physical identifier + reading time + reading location" data packet to the central BIM platform server via the construction site's wireless network (such as Wi-Fi or 4G / 5G).

[0068] After receiving the identification code data packet, the BIM platform server automatically matches it in its database. The platform maintains a pre-defined "object-identifier code" mapping table, recording the correspondence between all BIM model objects (virtual) and their planned assigned physical identification codes. When an identification code from the physical world is received, the system searches this table. If a completely matching record is found, the match is confirmed, and a record is created or updated in another "dynamic mapping relationship table." This record contains the physical identification code, the corresponding BIM model object ID, the timestamp of the most recent read, and the read location. This dynamic mapping relationship table is continuously called by subsequent data processing flows (such as step S2) as the basis for binding real-time sensor data streams to the correct BIM objects. If an unregistered identification code is read, the system will issue an abnormal alarm.

[0069] This implementation method clarifies a standardized and automated method for establishing physical-digital object identification mappings. In existing technologies, the association between construction machinery, vehicles, and components and BIM models often relies on manual recording, visual verification, or simple self-numbering. These methods are inefficient, error-prone, and struggle to support the real-time automatic collection of massive, dynamic data. This solution assigns physical objects a unified, machine-recognizable unique identifier and utilizes on-site identification equipment for automatic capture and uploading, achieving instantaneous and accurate matching between physical entities and the BIM virtual model when they enter the monitoring range. This mechanism fundamentally opens up the association channel between real-time IoT data streams and the BIM semantic model, providing a stable and reliable data association foundation for subsequent refined dynamic calculation of carbon footprints based on specific construction activities (rather than general equipment), thus solving the "information silo" problem.

[0070] In another embodiment of the present invention, the BIM-based method for dynamic monitoring of carbon footprint in modular steel structure construction, in step S22, according to the identifier mapping relationship, alignment and spatiotemporal fusion are performed within a unified operation time window. Specifically, a unified reference time server is deployed at the construction site to provide a unified time stamp for unit carbon footprint data packages and transportation status data. During data alignment, the data of the same physical object within a preset time tolerance range are fused based on the time stamp to generate a comprehensive data record with a unified time reference.

[0071] In this embodiment, a unified reference time server is deployed at the construction site. This server can be an industrial-grade time synchronization device supporting Network Time Protocol (NTP) or Precision Time Protocol (PTP), and its time source can be connected to the BeiDou Navigation Satellite System or Global Positioning System (GPS) signals to achieve millisecond-level time synchronization accuracy. This server can be installed in the computer room of the construction site command center and connected to the network switches throughout the site via an industrial Ethernet network. The edge computing unit can be an industrial IoT gateway with network time synchronization capabilities. It connects to the site's local area network via a network cable or wireless access point and is configured to automatically initiate time synchronization requests to the reference time server, with a synchronization period set to once every 30 seconds. Intelligent vehicle terminals installed on transport vehicles can have built-in communication modules supporting NTP client functionality, synchronizing time with the reference time server via a 4G or 5G network.

[0072] Unit carbon footprint data packets are generated by edge computing units. During packet encapsulation, the edge computing unit uses the locally synchronized system time to timestamp the data packets. The timestamp format can uniformly adopt "YYYY-MM-DD HH:MM:SS.sss". Transportation status data is generated and reported by vehicle terminals, which also use the synchronized terminal system time to timestamp the data during generation. After receiving all data, the central data processing platform first classifies the data packets according to the device identifiers in the data packets, and then aligns them based on the time stamps. For the same physical object (such as a crane), the system can set a preset time tolerance range, such as ±5 seconds. Timestamp data within this tolerance range is considered valid data within the same time window and then fused. During fusion, multiple unit carbon footprint data packets from the same device within the tolerance window can be sorted in chronological order, and their carbon emission equivalents can be accumulated. At the same time, the reported transportation status data (such as location and speed) within the same time period can be associated, ultimately generating a comprehensive data record with the start and end times after fusion as the markers, containing the cumulative carbon emissions and associated status information.

[0073] In existing construction carbon emission monitoring systems, various IoT devices typically rely on their own clocks to record data. Due to clock drift or lack of strict synchronization, timestamps from different sources often exhibit significant discrepancies. This makes it difficult to accurately determine which data belong to the same construction activity within the same spatiotemporal context when performing cross-device and cross-type data correlation analysis on the central platform, thus affecting the accuracy of carbon emission attribution and intensity calculation. This implementation method deploys a unified reference time server and enforces time synchronization for all data acquisition terminals across the site, ensuring that all real-time data has a unified and reliable time reference. By setting a clear time tolerance window for data alignment and fusion, the problem of data misalignment caused by transmission delays or minor differences in device clocks can be effectively solved. This allows the integrated data record generated by fusion to more accurately reflect the complete carbon footprint and operational status of a specific construction object within a specific time period, providing a time-consistent data foundation for subsequent accurate calculation of carbon emission intensity and improving the reliability and accuracy of dynamic monitoring.

[0074] In another embodiment of the present invention, the method for constructing and updating the dynamic carbon footprint factor correction model in step S212 of the BIM-based modular steel structure construction carbon footprint dynamic monitoring method is as follows: S2121. Based on historical data, establish a machine learning model for each type of construction machinery under different operating conditions, with key operating parameters as input and unit energy consumption correction coefficient as output. Key operating parameters include at least real-time load rate, engine speed, and working pressure. S2122. Input the current operating conditions and corresponding real-time operating parameters into the machine learning model and output the unit energy consumption correction coefficient. S2123. Multiply the baseline carbon footprint factor by the unit energy consumption correction coefficient to obtain the dynamic carbon footprint factor at the current calculation time. S2124. Regularly use newly added actual energy consumption and operating condition data as training samples to iteratively update the machine learning model.

[0075] In this embodiment, based on historical data, a machine learning model is established for each type of construction machinery under different operating conditions, with key operating parameters as input and a unit energy consumption correction coefficient as output. Key operating parameters may include real-time load rate, engine speed, and working pressure. Taking a truck crane as an example, the real-time load rate can be measured by a tension sensor installed at the fixed end of the hook wire rope, the engine speed can be read from the vehicle controller's local area network bus, and the working pressure can be obtained by a pressure sensor installed on the main lifting hydraulic circuit. The established machine learning model can use either gradient boosting regression tree or support vector regression algorithms. The training data for the model comes from historical datasets accumulated in past projects for this type of machinery. This dataset contains the aforementioned key operating parameters and rigorously measured actual unit-time energy consumption values ​​under different operating conditions. Before model training, the historical data can be cleaned, for example, by removing outliers caused by sensor malfunctions, and the parameters can be normalized. During training, 70% of the dataset can be used as the training set and 30% as the test set. The model hyperparameters are adjusted to minimize the mean square error between the predicted energy consumption correction coefficient and the actual value.

[0076] During real-time calculations, the system calls upon a baseline carbon footprint factor matching the current operating condition, which can be referenced from a nationally published emission coefficient database. Simultaneously, the currently identified operating condition and its corresponding real-time operating parameters are input into a pre-trained machine learning model. The model outputs a unit energy consumption correction coefficient, typically ranging from 0.5 to 1.5. The baseline carbon footprint factor is then multiplied by this correction coefficient to obtain the dynamic carbon footprint factor used for the current calculation moment. For example, when a crane is in a "lifting operation" condition, if the real-time load rate is high and the engine speed is stable, the model may output a correction coefficient less than 1.0, indicating that the mechanical energy efficiency is high and its carbon emission intensity per unit energy consumption is lower than the baseline level.

[0077] To maintain model accuracy, the system iteratively updates it periodically. The update process can be set to be triggered weekly or after each major construction milestone. During an update, newly collected actual energy consumption data from the past week, along with corresponding operating conditions and parameters, are cleaned and added to the historical dataset as new training samples. Subsequently, the expanded dataset is used to retrain or incrementally learn the original model. A validation step can be introduced during the update process, such as using recent data to verify the prediction accuracy of the updated model. If accuracy decreases, the reasons are analyzed and the training strategy is adjusted. In this way, the model can gradually adapt to the natural degradation of mechanical performance, differences in driver operating habits, and the influence of specific construction site environments, thereby continuously improving the accuracy and adaptability of dynamic carbon factor correction.

[0078] Existing methods for calculating machinery carbon emissions typically use fixed emission coefficients or simply switch between a limited number of preset operating conditions (such as idling and running), without considering the differences in energy efficiency caused by fluctuations in key operating parameters such as real-time load and machine status under the same operating condition. This implementation method, by establishing machine learning models for each type of machinery under different operating conditions, can dynamically and precisely correct energy consumption coefficients based on multi-dimensional real-time parameters such as load rate and speed, thereby obtaining a carbon footprint factor that more accurately reflects the instantaneous working state of the machinery. This shift from "fixed or coarsely segmented" to "based on multi-parameter dynamic learning" allows the underlying parameters for carbon equivalent conversion to more realistically reflect the actual operating energy efficiency of the machinery, thus improving the accuracy of individual machinery carbon footprint calculations from the source and providing a more reliable foundation for accurate monitoring of the overall project carbon footprint.

[0079] In another embodiment of the present invention, the BIM-based dynamic monitoring method for carbon footprint of modular steel structure construction, in step S3, the optimization target scheme issues new instructions through the following hierarchical mechanism: a) Decompose the optimization scheme into automated execution instructions and manually assisted execution instructions; b) For construction machinery connected to a controller and capable of automatic execution, the automatic execution command will be directly sent to its controller; c) For processes requiring manual intervention, push the manual assistance instructions to the mobile terminals of relevant personnel; d) All instructions are accompanied by a unified effective time and are simultaneously visualized in the BIM platform.

[0080] In this embodiment, after the optimized solution is generated, it is first deconstructed into specific executable instructions. The system can classify these instructions into two categories based on the type of resources they control and the execution method: automated execution instructions and manually assisted execution instructions. For example, for a self-propelled modular transport vehicle (SPMT) that can access a control network, the instruction to adjust its route can be deconstructed into an automated instruction containing a series of target coordinates and speed parameters. Conversely, for adjusting the lifting sequence requiring tower crane operator intervention, it is deconstructed into graphic information containing component numbers, suggested lifting timing, and precautions, serving as a manually assisted instruction. The deconstruction process relies on a pre-set instruction template library, which defines the instruction data formats acceptable to different types of machinery or processes.

[0081] The instruction issuance adopts a hierarchical mechanism. For construction machinery such as SPMTs and some intelligent concrete pump trucks that are connected to programmable logic controllers (PLCs) or dedicated controllers and have automatic execution capabilities, the construction management platform directly sends the encapsulated automated execution instruction data packets to the machine's controller via the on-site industrial wireless network. After receiving and verifying the instructions, the controller can store them in the execution queue. For processes requiring the intervention of personnel such as hoisters and signalmen, the system pushes the corresponding manual assistance instructions to the mobile terminals held by relevant personnel through the construction management platform's application programming interface (API). These terminals can be industrial explosion-proof tablets or smartphones with 4G / 5G capabilities, and the dedicated application on the terminal is responsible for receiving and displaying the instructions.

[0082] All instructions come with a unified effective time, determined by the system when generating optimization solutions, such as "6:00 AM the next day." This effective time is written into the metadata of each instruction. In the collaborative management view of the BIM platform, this is simultaneously visualized: next to the corresponding 3D model object (such as a transport vehicle or tower crane model) and on the timeline, the content of the new instruction, the target object, and the effective time are clearly marked with highlighted icons, text labels, or color changes. This provides site managers with a global and intuitive interface for monitoring instruction status.

[0083] Existing methods for communicating construction optimization plans rely heavily on meetings, paper documents, or simple mass messaging. These methods suffer from limited instruction formats and difficulty in ensuring synchronized understanding of the instructions, especially the precise effective time, among all implementers. This can easily lead to disconnects between automated systems and manual operations, and between different trades, causing waiting, conflicts, or misoperations, ultimately diminishing the expected effectiveness of the optimization plan. This implementation method deconstructs the plan into instructions adaptable to different implementers and employs a tiered distribution channel (direct connection to the controller and push to mobile terminals), ensuring the accuracy of the instruction format and the reliability of transmission. In particular, by attaching a unified effective time to all instructions and centrally visualizing it on the BIM platform, a clear and synchronized execution clock reference is created. This effectively coordinates the work rhythms of automated equipment and manual operations, ensuring that the optimization plan can be implemented collaboratively according to the predetermined sequence, thus improving the overall coordination and execution efficiency of construction adjustments.

[0084] Based on the same inventive concept, this invention also provides a BIM-based dynamic monitoring system for the carbon footprint of modular steel structure construction, comprising: The baseline generation module is used to integrate the 4D schedule, machinery resource library and transportation route of modular steel structure construction in the BIM platform, pre-associate carbon footprint factors for components, machinery and transportation tasks in the model, and generate a predicted carbon footprint baseline with time as the axis through construction simulation. A real-time monitoring and deviation calculation module is used to collect mechanical energy consumption and transportation data through IoT devices during construction, and dynamically bind the real-time data stream to the corresponding construction object based on the identifier mapping relationship between the BIM model and physical equipment; this module further includes: The edge computing unit is used to receive raw high-frequency data collected by the mechanical energy consumption sensor, perform data cleaning, operating condition identification, and preliminary carbon equivalent conversion based on the preset carbon footprint factor, and generate standardized unit carbon footprint data packages. The data fusion unit is used to perform spatiotemporal alignment and fusion of the unit carbon footprint data packets and transportation status data reported by each edge computing unit within a unified operation time window, based on the aforementioned identifier mapping relationship, to form a comprehensive data record. The deviation calculation unit is used to dynamically calculate the actual carbon emission intensity of each construction activity per unit time based on the comprehensive data records and pre-associated carbon footprint factors, and compare it with the predicted intensity at the corresponding time point in the predicted carbon footprint baseline in real time, and output the dynamic deviation. The optimization and instruction execution module is used to automatically initiate a construction scheme simulation optimization starting from the current state and covering the remaining project when the cumulative dynamic deviation exceeds a preset threshold and reaches a preset number of times; dynamically adjust the machinery scheduling, process sequence, or transportation scheme in the BIM-4D environment, simulate and calculate the carbon footprint effect after adjustment; and deconstruct the scheme that meets the optimization objective into a collaborative operation instruction set for issuance, while reconstructing the curve of the future time period in the predicted carbon footprint baseline.

[0085] The system may further include: an identification mapping management module, used to establish and maintain a dynamic mapping relationship between physical devices and BIM model objects through RFID or QR code tags; a digital twin simulation module, used to build a high-fidelity synchronous simulation environment for conflict pre-simulation and effect verification before the instructions are issued; and a dynamic carbon factor correction module, used to correct the benchmark carbon footprint factor in real time based on a machine learning model.

[0086] Based on the same inventive concept, the present invention also provides an electronic device, comprising: at least one processor, a memory, and a computer program stored in the memory and executable on the at least one processor, wherein the processor executes the computer program to implement the steps of the BIM-based modular construction carbon footprint dynamic monitoring method for steel structures as described above.

[0087] The electronic device may be an edge computing server deployed on the construction site, a central server in the project command center, or a cloud computing platform. The device may also include a sensor interface unit communicatively connected to the processor for accessing IoT devices such as mechanical energy consumption sensors, GPS terminals, and RFID readers; a network communication unit for data interaction with the BIM platform, mobile terminals, and mechanical controllers; and a time synchronization unit for receiving time signals from a reference time server and adding a unified timestamp to the data.

[0088] Based on the same inventive concept, the present invention also provides a computer-readable storage medium storing a computer program, which, when executed by a processor, implements the steps of the BIM-based dynamic monitoring method for carbon footprint of modular steel structure construction as described above.

[0089] The computer program can be loaded into the aforementioned electronic device or system to control the entire process of generating a carbon footprint baseline, collecting and fusing real-time data, dynamically calculating and analyzing carbon emission intensity, simulating and optimizing construction plans, and generating and issuing optimization instructions. The storage medium includes, but is not limited to, various media capable of storing program code, such as USB flash drives, external hard drives, ROM, RAM, magnetic disks, or optical disks.

[0090] The number of devices and processing scale described herein are for the purpose of simplifying the description of the invention. Applications, modifications, and variations of the invention will be readily apparent to those skilled in the art.

[0091] Although the embodiments of the present invention have been disclosed above, they are not limited to the applications listed in the specification and embodiments. They can be applied to various fields suitable for the present invention. For those skilled in the art, other modifications can be easily made. Therefore, without departing from the general concept defined by the claims and their equivalents, the present invention is not limited to the specific details.

Claims

1. A BIM-based method for dynamic monitoring of the carbon footprint of modular steel structure construction, characterized in that, include: S1. Integrate the 4D schedule, machinery resource library and transportation route for modular steel structure construction into the BIM platform, pre-associate carbon footprint factors for components, machinery and transportation tasks in the model, and generate a predicted carbon footprint baseline with time as the axis through construction simulation. S2. During construction, mechanical energy consumption and transportation data are collected through IoT devices, and based on the identifier mapping relationship between the BIM model and physical equipment, real-time data streams are dynamically bound to the corresponding construction objects; streaming computation is performed based on the bound data, specifically: S21. The edge computing unit at the construction site receives the raw high-frequency data collected by the mechanical energy consumption sensor, performs data cleaning, working condition identification, and preliminary carbon equivalent conversion based on the preset carbon footprint factor, and generates a standardized unit carbon footprint data package. S22. Align and spatiotemporally fuse the unit carbon footprint data packets, transportation paths, and status data of each edge computing unit with the identification mapping relationship within a unified operation time window to form a comprehensive data record. S23. Based on comprehensive data records and pre-correlated carbon footprint factors, dynamically calculate the actual carbon emission intensity per unit time for each construction activity, and compare it in real time with the predicted intensity at the corresponding time point in the predicted carbon footprint baseline, and output the dynamic deviation. S3. When the cumulative dynamic deviation exceeds the preset threshold and reaches the preset number of times, the BIM platform automatically starts the construction scheme simulation optimization with the current state as the starting point and the remaining project as the scope; dynamically adjusts the machinery scheduling, process sequence or transportation scheme in the BIM-4D environment, and simulates and calculates the carbon footprint effect after the adjustment; issues the scheme that meets the optimization goal as a new instruction, and simultaneously reconstructs the curve of the future period in the predicted carbon footprint baseline.

2. The method for dynamic monitoring of carbon footprint of modular steel structure construction based on BIM as described in claim 1, characterized in that, Step S21 involves identifying the operating conditions and performing preliminary carbon equivalent conversion based on preset carbon footprint factors to generate standardized unit carbon footprint data packages, specifically as follows: S211. A multi-source sensor group integrated on the construction machinery collects real-time energy consumption data and working condition data, and combines it with a pre-set mechanical operation mode classifier to identify the current working condition. Operating data includes attitude data, load data, and control signal data; S212. Call the benchmark carbon footprint factor that matches the current working condition, and use the dynamic carbon footprint factor correction model to correct the benchmark carbon footprint factor in real time to generate the dynamic carbon footprint factor; the input of the correction model includes at least the current working condition, real-time energy consumption data, and the benchmark value of energy consumption per unit time of the machine under the current working condition based on historical data. S213. Utilize the dynamic carbon footprint factor to perform carbon equivalent conversion on real-time energy consumption data and generate a unit carbon footprint data package, which includes at least the construction machinery identifier, timestamp, operating conditions, dynamic carbon footprint factor, and calculated carbon emission equivalent.

3. The method for dynamic monitoring of carbon footprint of modular steel structure construction based on BIM as described in claim 1, characterized in that, After outputting the dynamic deviation in step S23, the step also includes a deviation root cause tracing step: S231. Construct and maintain a multidimensional construction status dataset, which includes at least actual carbon emission intensity, 4D progress status, resource status and utilization rate of each machine, process sequence, and environmental monitoring data. S232. When the dynamic deviation exceeds the preset threshold, based on the causal rule base and correlation analysis algorithm, the dynamic deviation is spatiotemporally correlated and pattern matched with the abnormal variables in the multidimensional construction status dataset to identify the dominant variable causing the deviation. S233. Based on the dominant variables and their influence weights, generate a root cause analysis report and output the causes of carbon footprint deviation in probability order. The types of causes of carbon footprint deviation include uneconomical operation of machinery, idle resources caused by process conflicts, congestion of transportation routes, or discrepancies between actual construction conditions and model presets.

4. The method for dynamic monitoring of carbon footprint of modular steel structure construction based on BIM as described in claim 1, characterized in that, In step S3, before issuing a new instruction based on the solution that meets the optimization objective, the following steps are also included: S31. Using the current BIM model status, the real-time site status collected by the Internet of Things, and the optimization scheme as inputs, construct a high-fidelity synchronous simulation environment in the digital twin layer. S32. Accelerate the pre-execution optimization scheme in a synchronous simulation environment, simultaneously simulate construction progress, resource movement and carbon emissions, and detect whether there are resource space conflicts, path conflicts or violations of safety rules during the pre-execution process. S33. If the pre-run passes and the carbon footprint optimization effect is verified, the optimization scheme will be deconstructed into a set of cooperative operation instructions with time constraints. S34. Distribute the collaborative operation instruction set to the corresponding mechanical controllers or personnel mobile terminals through the construction management platform to ensure that the instructions take effect and are executed after the set unified time reference point.

5. The method for dynamic monitoring of carbon footprint of modular steel structure construction based on BIM as described in claim 1, characterized in that, The identification mapping relationship is established in the following way: RFID or QR code tags containing unique model identification codes are assigned to construction machinery, transport vehicles and prefabricated components involved in the construction; after the on-site identification equipment reads the tag information, it automatically matches it with the identification code of the corresponding object in the BIM model to form and maintain a dynamic mapping relationship table.

6. The method for dynamic monitoring of carbon footprint of modular steel structure construction based on BIM as described in claim 1, characterized in that, In step S22, based on the identifier mapping relationship, alignment and spatiotemporal fusion are performed within a unified operation time window. Specifically, a unified reference time server is deployed at the construction site to provide a unified time stamp for unit carbon footprint data packages and transportation status data. During data alignment, the data of the same physical object within a preset time tolerance range are fused based on the time stamp to generate a comprehensive data record with a unified time reference.

7. The method for dynamic monitoring of carbon footprint of modular steel structure construction based on BIM as described in claim 2, characterized in that, The method for constructing and updating the dynamic carbon footprint factor correction model in step S212 is as follows: S2121. Based on historical data, establish a machine learning model for each type of construction machinery under different operating conditions, with key operating parameters as input and unit energy consumption correction coefficient as output. Key operating parameters include at least real-time load rate, engine speed, and working pressure. S2122. Input the current operating conditions and corresponding real-time operating parameters into the machine learning model and output the unit energy consumption correction coefficient. S2123. Multiply the baseline carbon footprint factor by the unit energy consumption correction coefficient to obtain the dynamic carbon footprint factor at the current calculation time. S2124. Regularly use newly added actual energy consumption and operating condition data as training samples to iteratively update the machine learning model.

8. The method for dynamic monitoring of carbon footprint of modular steel structure construction based on BIM as described in claim 1, characterized in that, In step S3, the optimization target scheme is used to issue new instructions through the following hierarchical mechanism: a) Decompose the optimization scheme into automated execution instructions and manually assisted execution instructions; b) For construction machinery connected to a controller and capable of automatic execution, the automatic execution command will be directly sent to its controller; c) For processes requiring manual intervention, push the manual assistance instructions to the mobile terminals of relevant personnel; d) All instructions are accompanied by a unified effective time and are simultaneously visualized in the BIM platform.