A method for accounting for carbon emissions in the whole life cycle of an expressway

Abstract

This invention discloses a method for statistical accounting of carbon emissions throughout the entire life cycle of highways, belonging to the field of carbon emission accounting technology. The method includes: dividing the entire life cycle of a highway into multiple stages and dynamically defining the accounting boundaries of each stage based on spatiotemporal dimensions; setting carbon emission accounting indicators for all types and collecting multimodal data corresponding to each stage; building a hierarchical computing architecture based on edge-cloud collaborative technology to calculate the carbon emissions of each stage, and using a virtual highway model that can learn and update online to verify the carbon emission accounting results; and drawing a carbon footprint heatmap based on the carbon emission accounting results to display the spatiotemporal evolution process. This invention covers the construction, operation, and demolition stages of the entire highway life cycle, ensuring the comprehensiveness and systematic nature of the accounting results; employing multiple data sources and collection methods, combined with corresponding accounting methods, ensures the accuracy and reliability of the accounting results, providing a scientific basis for the low-carbon construction and operation of highways.

Description

Technical Field

[0001] This invention relates to the field of carbon emission accounting technology, and in particular to a method for statistical accounting of carbon emissions throughout the entire life cycle of highways. Background Technology

[0002] As an important transportation infrastructure, the carbon emissions generated by highways throughout their entire life cycle (construction, operation, and demolition) cannot be ignored.

[0003] Currently, various methods exist for carbon emission accounting both domestically and internationally, such as the emission factor method, mass balance method, and measurement method. These methods have been widely applied across different sectors and emission source types. However, for the specific sector of highways, a complete and systematic method for calculating carbon emissions throughout their entire life cycle has not yet been established. Existing methods often focus on carbon emission accounting at a single stage, neglecting the characteristics and interrelationships of carbon emissions at different stages throughout the highway's life cycle, resulting in inaccurate and incomplete accounting results.

[0004] Existing methods for calculating carbon emissions from highways suffer from the following shortcomings: 1. Unclear accounting boundaries: Existing methods often focus only on carbon emissions during a specific phase of the highway's lifecycle, neglecting the statistics of carbon emissions throughout the entire lifecycle. 2. Limited data sources: These methods rely heavily on theoretical calculations or limited actual monitoring data, lacking comprehensive and systematic data support. 3. Inconsistent accounting methods: Significant differences exist between different methods, resulting in a lack of comparability and accuracy in the accounting results. Summary of the Invention

[0005] Therefore, it is necessary to provide a method for statistical accounting of carbon emissions throughout the entire life cycle of highways to address the aforementioned technical issues.

[0006] In a first aspect, the present invention provides a method for statistical accounting of carbon emissions throughout the entire life cycle of highways, including:

[0007] S1. Divide the entire life cycle of the expressway into multiple stages, and dynamically define the accounting boundaries of each stage based on the spatiotemporal dimension.

[0008] S2. Set up carbon emission accounting indicators for all types of carbon emissions and collect multimodal data for each stage.

[0009] S3. Build a layered computing architecture based on edge-cloud collaborative technology to calculate carbon emissions at each stage, and use a virtual highway model that can be learned and updated online to verify the carbon emission calculation results.

[0010] S4. Draw a carbon footprint heat map based on carbon emission accounting results to show the spatiotemporal evolution process.

[0011] Furthermore, the entire lifecycle of a highway is divided into multiple stages, and the accounting boundaries of each stage are dynamically defined based on spatiotemporal dimensions, including:

[0012] S11. The entire life cycle of the expressway is divided into the construction phase, the operation phase, and the demolition phase, and the duration of each phase is dynamically adjusted according to the actual progress of the project.

[0013] S12. Utilize highway engineering information in conjunction with geographic information systems and building information models to define the physical impact range and emission responsibility areas for carbon emissions at each stage.

[0014] S13. Construct a carbon flow transfer matrix between each stage to quantify the transfer effect of carbon emissions at each stage.

[0015] Furthermore, the duration of each stage will be dynamically adjusted based on the actual progress of the project, including:

[0016] S111. Obtain the building information model and engineering machinery data of the highway project, divide the highway construction area into several unit grids, and generate a machinery coverage matrix.

[0017] S112. Compare the actual coverage grid of the highway with the design grid, calculate the real-time progress, and use a long short-term memory network to predict the future construction progress. When the construction progress exceeds the preset progress threshold, switch the highway from the construction phase to the operation phase and mark the construction and operation time information.

[0018] S113. Obtain the pavement performance index of the expressway. When the pavement performance index is lower than the preset performance threshold, switch the expressway from the operation stage to the demolition stage and mark the demolition time information.

[0019] Furthermore, a layered computing architecture based on edge-cloud collaborative technology is built to calculate carbon emissions at each stage, and the carbon emission calculation results are verified using a virtual highway model that can be learned and updated online.

[0020] S31. Build a layered computing architecture consisting of an edge layer, a fog layer, and a cloud layer, and deploy it sequentially on the engineering machinery end, the road section management center, and the traffic control platform to realize full life cycle computing of highways;

[0021] S32. Calculate the total carbon emissions during the construction phase using a hierarchical computing architecture;

[0022] S33. Calculate the total operational carbon emissions at the road network level during the operational phase using a hierarchical computing architecture.

[0023] S34. Calculate the total carbon emissions during the demolition phase using a hierarchical computing architecture and integrate them throughout the entire life cycle.

[0024] S35. Construct a virtual highway model based on digital twins, introduce an incremental online learning mechanism, integrate the computational data of each layer in the hierarchical computing architecture, and realize cross-stage and cross-temporal coupled simulation.

[0025] Furthermore, the calculation of total construction carbon emissions during the construction phase using a hierarchical computing architecture includes:

[0026] S321. Divide the multimodal data of the construction phase into different construction emission types and extract the corresponding accounting indicators. The construction emission types include equipment operation, building material production and transportation.

[0027] S322. Calculate the single-unit carbon emissions of each construction emission type during the construction phase using the edge layer;

[0028] S323. Utilize fog layers to unify the mechanical trajectories of construction machinery with the coordinates of the building information model during the construction phase, calibrate the clocks of various equipment, and integrate and calculate the total carbon emissions of the construction phase.

[0029] S324. Utilize the cloud layer to match the factor library of engineering machinery and building materials, dynamically update the edge layer and fog layer, and monitor the abnormal accounting data of the edge layer and fog layer in real time.

[0030] Furthermore, the total operational carbon emissions at the road network level during the operational phase are calculated using a hierarchical computing architecture, including:

[0031] S331. Divide the multimodal data of the operation phase into different operation emission types and extract the corresponding accounting indicators. The operation emission types include vehicle type and facility maintenance type.

[0032] S332. Use the edge layer to preprocess the multimodal data and extract the non-stop toll collection data of vehicles, mobile phone signaling data and road monitoring data from various accounting indicators;

[0033] S333. By integrating electronic non-stop toll collection data and mobile phone signaling data using fog layers, a minute-level traffic flow matrix is ​​generated, traffic flow carbon emissions are calculated, and combined with the calculated road maintenance carbon emissions, the total operating carbon emissions of a single expressway during its operation phase are generated.

[0034] S334. Utilize cloud layer for micro-traffic simulation, simulate the correlation curve between congestion index and emissions, and calculate the carbon emissions of the entire road network in parallel through road network-level integration.

[0035] Furthermore, a hierarchical computing architecture is used to calculate the total carbon emissions during the demolition phase, and a full lifecycle integration is performed, including:

[0036] S341. Divide the multimodal data of the demolition phase into different demolition emission types and extract the corresponding accounting indicators. The demolition emission types include demolition equipment type and waste treatment type.

[0037] S342. Use the edge layer to identify and record data on waste on highways and usage data of dismantling equipment;

[0038] S343. Use fog to generate a unique digital ID for each batch of recycled waste, record the transportation distance, processing technology and reuse projects of the waste, and calculate the total carbon emissions of demolition during the demolition phase.

[0039] S344. Utilize the cloud layer to perform circular economy offset calculations, integrate and calculate the carbon emission distribution and total carbon emissions throughout the entire life cycle of highways, store the output results of each layer in the cloud, and record the source, processing path and quality label of each data through data lineage tracking.

[0040] Furthermore, a virtual highway model based on digital twins is constructed, an incremental online learning mechanism is introduced, and computational data from each layer within the hierarchical computing architecture are integrated to achieve coupled simulation across stages and across time and space, including:

[0041] S351. Combining geographic information systems, building information models, and multiphysics coupling, a virtual highway model based on digital twins is constructed, and the output data of the hierarchical computing architecture is integrated.

[0042] S352. Embed an incremental online learning module in the virtual highway model, regularly update edge and cloud data, and adjust and correct the model parameters in the virtual highway model.

[0043] S353. Use a virtual highway model to perform coupled simulations of different stages, regions, and time periods throughout the entire life cycle of a highway, and simulate the interrelationships between different stages.

[0044] S354. By constructing a highway knowledge graph, the virtual highway model is adapted for general use.

[0045] Furthermore, a virtual highway model is used to conduct coupled simulations of different stages, regions, and time periods throughout the entire lifecycle of a highway, simulating the interrelationships between different stages, including:

[0046] S3531. Establish state vectors for the construction phase, operation phase, and demolition phase respectively;

[0047] S3532. Introduce a multi-scale simulation module into the virtual highway model and use a spatiotemporal coupling simulation algorithm to interactively fuse multimodal data and calculation results from each stage and region.

[0048] S3533. The multi-scale simulation modules at each stage transmit data through a unified interface and adopt a hierarchical simulation strategy to integrate local short-term dynamics with global long-term trends, thereby achieving cross-temporal and spatial simulation.

[0049] S3534. The carbon emission data output by the hierarchical computing architecture is used as the actual accounting value and compared with the carbon emission prediction value of the virtual highway model coupled simulation. The difference between the prediction value and the actual accounting value is quantitatively evaluated, and the difference value is fed back to the incremental online learning module to optimize the parameters of the virtual highway model.

[0050] Furthermore, by constructing a highway knowledge graph, the virtual highway model can be generalized and adapted, including:

[0051] S3541. Set up a unified feature space, use a pre-trained model as a feature extractor, freeze the underlying network of the virtual highway model, and build a transfer learning framework.

[0052] S3542. Construct a highway knowledge graph to store prior knowledge about highways, and when a new project is input, automatically match similar historical cases to initialize the model parameters of the virtual highway model.

[0053] The beneficial effects of this invention are as follows:

[0054] 1. This invention covers the construction, operation, and demolition stages of the entire life cycle of highways, ensuring the comprehensiveness and systematic nature of the accounting results; it employs multiple data sources and collection methods, combined with corresponding accounting methods, to ensure the accuracy and reliability of the accounting results; at the same time, it provides detailed accounting processes and steps, facilitating practical operation and application, and providing a scientific basis for the low-carbon construction and operation of highways.

[0055] 2. By dividing the entire life cycle of highways into construction, operation, and demolition phases, and dynamically defining the accounting boundaries of each phase based on spatiotemporal dimensions, refined carbon emission management is achieved throughout the entire process and at all times. The use of dynamic boundary correction technology can accurately capture project progress and environmental changes, ensuring the real-time nature and representativeness of carbon emission data at each phase. This breaks through the limitations of traditional methods, such as single phase division and incomplete data collection, and provides a scientific basis for green construction and low-carbon operation. It not only significantly improves the accuracy and credibility of the accounting results, but also provides data support for subsequent carbon emission reduction strategy optimization and environmental impact assessment.

[0056] 3. By setting all types of carbon emission indicators and collecting multimodal data corresponding to each stage, efficient integration of various types of information is achieved. With the help of a layered computing architecture that integrates edge and cloud, on-site sensing and historical data can be processed in real time, and data fusion algorithms can be used to eliminate the differences caused by a single accounting method. At the same time, the online learning and updating virtual highway model dynamically verifies and self-corrects the calculation results, ensuring that the accounting results at each stage reach a high level in terms of accuracy and comparability. This effectively reduces data noise and errors and improves the scientificity and adaptability of the whole life cycle carbon emission accounting. Attached Figure Description

[0057] The accompanying drawings, which are included to provide a further understanding of the invention and form part of this invention, illustrate exemplary embodiments of the invention and are used to explain the invention, but do not constitute an undue limitation of the invention. In the drawings:

[0058] Figure 1 This is a flowchart of a method for statistical accounting of carbon emissions throughout the entire life cycle of a highway, according to an embodiment of the present invention. Detailed Implementation

[0059] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention.

[0060] Please see Figure 1 This paper provides a method for statistical accounting of carbon emissions throughout the entire life cycle of highways, including:

[0061] S1. Divide the entire life cycle of the highway into multiple stages, and dynamically define the accounting boundaries of each stage based on the spatiotemporal dimension.

[0062] In the description of this invention, the entire life cycle of a highway is divided into multiple stages, and the accounting boundaries of each stage are dynamically defined based on spatiotemporal dimensions, including:

[0063] S11. The entire life cycle of the expressway is divided into the construction phase, the operation phase, and the demolition phase, and the duration of each phase is dynamically adjusted according to the actual progress of the project.

[0064] Specifically, this invention divides the entire life cycle of a highway into a construction phase, an operation phase, and a demolition phase, and determines the accounting boundaries for each phase, including the following aspects:

[0065] 1. Construction phase: This includes the design, construction, and installation processes, and the accounting boundaries cover the operation of construction equipment, the production and transportation of building materials.

[0066] 2. Operation phase: This includes daily maintenance, upkeep, and usage. The accounting scope covers the operation of vehicles, the maintenance and upkeep of facilities, etc.

[0067] 3. Demolition Phase: This includes the demolition of facilities and waste disposal, and the accounting boundary covers the demolition equipment and waste disposal.

[0068] In the description of this invention, dynamically adjusting the duration of each stage according to the actual progress of the project includes:

[0069] S111. Obtain the building information model and engineering machinery data of the highway project, divide the highway construction area into several unit grids, and generate a machinery coverage matrix.

[0070] Specifically, a BIM (Building Information Modeling) system is used to extract detailed spatial information about the construction area, including road layout, component locations, and material arrangement. Simultaneously, location information and operational status data are collected in real time from sensors on construction machinery (such as excavators, bulldozers, and cranes). Based on the actual dimensions and construction characteristics of the area, the entire construction area is divided into several uniform grid cells. Each grid cell represents a minimum construction unit, facilitating subsequent monitoring and progress analysis.

[0071] The coverage of each grid by each machine at different time periods is statistically analyzed to form a machine coverage matrix. Each element in the matrix represents the machine operation status of the corresponding grid at a specific time, providing a quantitative basis for real-time assessment of construction progress.

[0072] S112. Compare the actual coverage grid of the highway with the design grid, calculate the real-time progress, and use a long short-term memory network to predict the future construction progress. When the construction progress exceeds the preset progress threshold, switch the highway from the construction phase to the operation phase and mark the construction and operation time information.

[0073] Specifically, by comparing the actual construction coverage grid with the grid layout required by the design, the current construction progress is calculated, and a Long Short-Term Memory (LSTM) network is used to predict the future construction progress. The specific implementation process is as follows:

[0074] 1. Progress Comparison Analysis: Compare the actual mechanical coverage matrix obtained in S111 with the pre-designed ideal coverage matrix, and calculate the ratio of the actual number of covered cells to the total number of designed coverage cells to obtain real-time construction progress indicators.

[0075] 2. Schedule Prediction: The LSTM model is trained on historical schedule data to capture the long-term dependencies and trends in schedule changes. The model input is coverage data over a continuous time period, and the output is the predicted construction schedule for a future period.

[0076] 3. Phase Switching Judgment: When the real-time or predicted construction progress exceeds the preset progress threshold, the system automatically determines that the construction phase of the project has been basically completed. At this time, a phase switching is triggered, changing the current state from "Construction Phase" to "Operation Phase". Simultaneously, the system records the time information of the end of the construction phase and the start of the operation phase to ensure accurate time data marking.

[0077] S113. Obtain the pavement performance index of the expressway. When the pavement performance index is lower than the preset performance threshold, switch the expressway from the operation stage to the demolition stage and mark the demolition time information.

[0078] Specifically, by acquiring pavement performance indices, the operational status of the highway is monitored, and it is determined whether the demolition phase is necessary. The specific implementation process is as follows:

[0079] 1. Performance Index Collection: Real-time collection of highway pavement performance indicators, such as smoothness, crack rate, and wear degree, using on-site testing equipment or remote sensing monitoring systems. Data processing generates a unified pavement performance index.

[0080] 2. Performance Threshold Comparison: The current pavement performance index is compared with a preset performance threshold. When the index is lower than the threshold, it indicates that the pavement has suffered significant damage or poses a safety hazard.

[0081] 3. Phase Switching and Time Marking: When the pavement performance index is detected to be lower than the preset value, the system automatically determines that the operation phase has ended and triggers the transition from the operation phase to the demolition phase. Simultaneously, the start time of the demolition phase is recorded to provide a time marker for subsequent demolition operations and carbon emission accounting.

[0082] S12. Utilize highway engineering information combined with geographic information systems and building information models to define the physical impact range and emission responsibility areas for carbon emissions at each stage.

[0083] Specifically, the project management system collects detailed design data, construction records, and operation and management information for highway projects, including road alignment, structural composition, material lists, construction techniques, and operation and maintenance plans. Simultaneously, a 3D building information model is extracted from the BIM system to obtain information on engineering components, equipment layout, and material usage. A GIS platform is used to collect geospatial data of the project area, such as topography, land use, ecological environment, transportation network, and administrative divisions. These multi-source data are then aligned spatially and temporally to form a unified data foundation.

[0084] Seamlessly integrating BIM models with GIS (Geographic Information System) platforms enables information complementarity through spatial overlay analysis. For example, by merging component locations in BIM with geographic locations, environmentally sensitive areas, and administrative boundaries in GIS, the spatial distribution of construction areas, equipment installation areas, and operational facilities can be clearly identified, thereby determining the specific physical impact range of activities at each stage.

[0085] During the construction phase, the integration of BIM and GIS delineates areas such as the construction site, material storage areas, equipment operation areas, and transportation corridors. These areas directly generate carbon emissions from machinery operation, energy consumption, and material transportation. During the operation phase, based on traffic flow, energy consumption data, and the distribution of on-site facilities, GIS spatial analysis is used to extract the emission impact areas of vehicle exhaust, lighting, and power supply systems. During the demolition phase, the carbon emission diffusion range is determined based on the demolition work area and waste disposal site. Through spatial overlay and buffer zone analysis, the physical impact range of carbon emissions at each stage is visually presented graphically.

[0086] Based on the project's organizational structure and management division of labor, the defined physical impact area is matched with the actual management responsibilities. Using a GIS platform, the administrative divisions of the project location and the internal zoning information of the project are overlaid to clarify the emission responsibility entities (such as construction units, operation and management units, and demolition and disposal units) in each area, forming a carbon emission accounting area with clearly defined responsibilities and corresponding rights and responsibilities.

[0087] S13. Construct a carbon flow transfer matrix between each stage to quantify the transfer effect of carbon emissions at each stage.

[0088] Specifically, constructing a carbon transfer matrix requires identifying the main carbon emission sources in each stage (construction, operation, and demolition). For example, the construction stage includes carbon emissions from construction machinery operation, material production, and transportation; the operation stage involves carbon emissions from daily energy consumption, vehicle exhaust, and maintenance; and the demolition stage focuses on carbon emissions from equipment dismantling, waste disposal, and resource recycling. Through on-site monitoring, historical data, and engineering models, the total direct and indirect emissions for each stage are calculated, providing a quantitative basis for constructing the transfer matrix.

[0089] Analyze the interactions between different phases; for example, carbon embedded in materials during the construction phase may be released during the operation phase through maintenance and reuse, or some of the embedded carbon may be recovered during the demolition phase, thus reducing overall emissions. Construct a matrix M, where each element M ij This represents the proportion of carbon emissions transferred from stage i to stage j. The transfer coefficient can be expressed by the formula:

[0090] ;

[0091] In the formula, Ei ΔE represents the total carbon emissions in stage i; ij The carbon emissions from stage i to stage j are represented; each transfer coefficient is determined by fitting using statistical analysis, regression models, or machine learning methods (such as random forests or neural networks).

[0092] A multi-source data acquisition approach was adopted, utilizing engineering monitoring data, real-time sensor data, and simulation data to obtain the actual emissions and transfer effects at each stage. Through data fusion and error correction, the transfer coefficients were dynamically adjusted to ensure the matrix reflects the true operational status of the project. Furthermore, cross-validation and historical data review were used periodically to calibrate and update the transfer matrix.

[0093] Finally, carbon emission data from each stage are coupled with the transfer matrix to simulate the overall carbon flow process. For example, matrix multiplication is used to calculate the cumulative carbon emissions at each stage throughout the entire life cycle, thus achieving closed-loop accounting of carbon emissions across the entire life cycle. This matrix not only quantifies the direct transfer effects between stages but also identifies errors caused by double counting or omissions, ensuring high accuracy and comparability of the accounting results.

[0094] S2. Set up carbon emission accounting indicators for all types and collect multimodal data corresponding to each stage.

[0095] In the description of this invention, the acquisition of multimodal data at each stage includes the following aspects:

[0096] 1. Data collection during the construction phase

[0097] 1.1 Intelligent construction machinery: Equipped with Beidou terminal + 5G module, it transmits fuel consumption data in real time (direct reading from CAN bus, accuracy ±1.2%) and operating trajectory (generating carbon emission intensity heat map).

[0098] 1.2 BIM Model Analysis: Extract building material usage through IFC standards (reinforcement / concrete error <2%).

[0099] 2. Operational monitoring network

[0100] 2.1 ETC Enhanced System (Electronic Toll Collection System): AI coprocessor (FPGA acceleration) is deployed on the gantry to realize vehicle type recognition (ResNet-50 model, accuracy >97%) and instantaneous emission calculation (improved COPERT model, resolution 0.1 seconds).

[0101] 2.2 Distributed Fiber Optic Sensing: DAS nodes (distributed antenna system nodes) are deployed every 500m to monitor axle load (to calculate rolling resistance coefficient) and road surface deformation (to predict emissions caused by maintenance needs).

[0102] 3. Demolition Period Traceability System

[0103] 3.1 UAV swarm modeling: Using UAVs equipped with LiDAR and SLAM algorithms to reconstruct the demolition site (point cloud density > 2000 points / m²), and using deep learning to segment waste types (concrete / asphalt classification accuracy > 93%).

[0104] 3.2 Blockchain Carbon Passport: Utilizes Hyperledger Fabric (an open-source blockchain distributed ledger) to record on-chain material crushing energy consumption (smart meter data uploaded to the blockchain) and the transportation trajectory of recycled materials (interfacing with freight platform APIs).

[0105] S3. Build a layered computing architecture based on edge-cloud collaborative technology to calculate carbon emissions at each stage, and use a virtual highway model that can be learned and updated online to verify the carbon emission calculation results.

[0106] In the description of this invention, a hierarchical computing architecture based on edge-cloud collaborative technology is constructed to calculate carbon emissions at each stage, and the carbon emission calculation results are verified using a virtual highway model that can be learned and updated online.

[0107] S31. Build a layered computing architecture consisting of an edge layer, a fog layer, and a cloud layer, and deploy them sequentially on the engineering machinery, the road section management center, and the traffic control platform to realize full lifecycle computing for highways.

[0108] Specifically, based on a layered computing architecture that integrates edge, fog, and cloud, this architecture is deployed sequentially at the engineering machinery end, the road section management center, and the traffic control platform, thereby enabling real-time calculation and monitoring of key indicators such as carbon emissions and energy consumption throughout the entire life cycle of highways (construction, operation, and demolition).

[0109] 1. Edge layer deployment (engineering machinery end):

[0110] Sensors, GPS devices, energy consumption monitoring devices, and other terminal equipment are deployed at construction sites and on construction machinery to collect real-time data on the machinery's operating status, fuel consumption, and equipment location. This data directly reflects the dynamic information of the construction site and forms the basis for subsequent data processing and carbon emission accounting. The edge layer has preliminary data cleaning and preprocessing functions, capable of formatting and denoising the raw data, and sending the data to the upper-layer platform through secure protocols.

[0111] 2. Fog Deployment (Road Section Management Center):

[0112] A fog computing platform is built at the road section management center to serve as an intermediate aggregation and processing node for edge data. The fog layer mainly undertakes the aggregation, rapid response, and preliminary analysis of local data. It further integrates preprocessed data from various engineering machinery terminals to perform local event detection, real-time monitoring, and some simple carbon emission calculations, thereby reducing data transmission latency and improving the emergency response capability within the area.

[0113] 3. Cloud-based deployment (traffic control platform):

[0114] A cloud computing center is deployed on the traffic control platform to comprehensively integrate and deeply analyze data from various road management centers using high-performance computing resources and a big data platform. The cloud layer performs full lifecycle model calculations, employing various accounting algorithms, data fusion, and simulation methods to conduct comprehensive carbon emission and energy consumption analysis of the construction, operation, and demolition phases of highways, and supports long-term trend prediction and dynamic optimization.

[0115] Through real-time data acquisition at the edge layer, local data processing at the fog layer, and deep integration and computing at the cloud, the entire layered architecture enables precise monitoring and accounting of the entire life cycle of highways, providing real-time and accurate data support and decision-making basis for low-carbon construction and operation management.

[0116] S32. Calculate the total carbon emissions of the construction phase using a hierarchical computing architecture.

[0117] In the description of this invention, calculating the total construction carbon emissions during the construction phase using a hierarchical computing architecture includes:

[0118] S321. Divide the multimodal data of the construction phase into different construction emission types and extract the corresponding accounting indicators. The construction emission types include equipment operation, building material production and transportation.

[0119] S322. Calculate the single-unit carbon emissions of each construction emission type during the construction phase using the edge layer.

[0120] Specifically, during construction, various types of machinery (such as excavators, pavers, and bulldozers) consume fuel or electricity during operation, and the combustion of fuel or the use of electricity leads to carbon emissions. Therefore, the carbon emissions of construction equipment are primarily calculated based on its energy consumption. For fuel-powered equipment, carbon emissions depend on the equipment's fuel consumption, including fuel type, equipment power, operating time, and load conditions. Generally, fuel consumption differs under full load, half load, and no-load conditions; therefore, the impact of load rate needs to be comprehensively considered to ensure the accuracy of the calculation.

[0121] For electrically driven equipment, the power consumption of the equipment must be considered, and the carbon emission factor of the local power grid must be taken into account. Since the energy structure of power grids varies across regions (e.g., using thermal power, wind power, or hydropower), the carbon emissions per unit of electricity also differ. Therefore, calculations need to be made in accordance with local power data. The carbon emissions of construction equipment mainly depend on the characteristics of fuel combustion. For fuel-fired equipment, emissions are related to the combustion efficiency and chemical composition of the fuel; different types of fuel (such as diesel and gasoline) produce different amounts of carbon dioxide emissions during combustion. While electrically driven equipment does not directly generate carbon emissions, the electricity it uses comes from the power grid, and the power generation method of the grid (e.g., coal power, nuclear power, wind power, etc.) determines its carbon emission level. Therefore, the carbon emission calculations for electrically driven equipment need to be adjusted according to the local power structure.

[0122] Furthermore, in actual operation, equipment does not always operate at its rated power but is affected by load variations. Therefore, a load factor needs to be incorporated into the calculation to reflect the actual fuel or electrical energy consumption of the equipment under different operating conditions, thereby improving the accuracy of the calculation.

[0123] The production of building materials is a significant source of carbon emissions from highway construction, especially high-energy-consuming materials such as cement, steel bars, and asphalt. Carbon emission accounting primarily considers the consumption of materials and the carbon emission factors of the production process. By statistically analyzing the various building materials required during construction and combining this data with the carbon emission data for each material's production, the total carbon emissions can be calculated.

[0124] Carbon emission factors for building materials are typically derived from Life Cycle Assessment (LCA) databases or industry statistics. For example, cement production involves high-temperature calcination, releasing large amounts of carbon dioxide; steel smelting also consumes significant amounts of energy. Therefore, the emission levels from the production of different materials vary considerably and need to be calculated based on the specific material type.

[0125] Carbon emissions from building materials primarily originate from the extraction, processing, energy consumption, and chemical reactions during production. For example, the calcination of limestone at high temperatures during cement production releases carbon dioxide, while steel smelting involves coal combustion and oxidation reactions. Therefore, each building material has a specific carbon emission factor that can be used to estimate the carbon emission levels of its production process.

[0126] S323. Utilize fog layers to unify the mechanical trajectories of construction machinery with the coordinates of the building information model during the construction phase, calibrate the clocks of various equipment, and integrate and calculate the total carbon emissions of the construction phase.

[0127] S324. Utilize the cloud layer to match the factor library of engineering machinery and building materials, dynamically update the edge layer and fog layer, and monitor the abnormal accounting data of the edge layer and fog layer in real time.

[0128] S33. Calculate the total operational carbon emissions at the road network level during the operational phase using a hierarchical computing architecture.

[0129] In the description of this invention, calculating the total operational carbon emissions at the road network level during the operational phase using a hierarchical computing architecture includes:

[0130] S331. Divide the multimodal data of the operation phase into different operation emission types and extract the corresponding accounting indicators. The operation emission types include vehicle type and facility maintenance type.

[0131] S332. Use the edge layer to preprocess the multimodal data and extract the non-stop toll collection data of vehicles, mobile phone signaling data and road monitoring data from various accounting indicators.

[0132] Specifically, the edge layer is deployed at ETC gantries and road monitoring nodes, and uses embedded AI chips (such as NVIDIA Jetson Xavier) to process multimodal data in real time: the ETC gantry runs a lightweight YOLOv5 model to classify vehicle types (accuracy > 98%), and calculates the single vehicle travel time and instantaneous speed by combining the license plate recognition results with the transaction timestamp.

[0133] Mobile signaling data is accessed through 5G CPE (terminal equipment) devices, and a spatial heat map is generated using a kernel density estimation algorithm to supplement the traffic flow distribution in road sections not covered by ETC. The strain waveform data of the road monitoring fiber is denoised by wavelet transform and then input into a pre-trained neural network model (64 nodes in the input layer and 32 nodes in the hidden layer) to calculate the vehicle axle load and rolling resistance coefficient. The edge nodes finally output structured data packets (containing fields such as timestamp, location coordinates, vehicle type, speed, and axle load), which are encrypted and transmitted to the fog layer via the MQTT protocol.

[0134] S333. By integrating electronic non-stop toll collection data and mobile phone signaling data using fog layers, a minute-level traffic flow matrix is ​​generated, and traffic flow carbon emissions are calculated. Combined with the calculated road maintenance carbon emissions, the total operational carbon emissions of a single highway during its operation phase are generated.

[0135] Specifically, at the road segment level fog calculation node, the ETC and mobile phone signaling data are first spatiotemporally aligned: based on the Kalman filter algorithm, the two types of data sources are fused to eliminate device clock deviation (error < 1 second) and positioning drift (spatial error < 10 meters) and generate a minute-level traffic flow matrix (dimension: time × road segment × vehicle type × speed range).

[0136] Then, the dynamic emission factor library (synchronizing grid carbon intensity data hourly) is called, and the improved MOVES model is used to calculate the carbon emission intensity of each cell. The formula is as follows:

[0137] ;

[0138] In the formula, N veh Indicates the number of vehicles, EF speed Indicates a rate-dependent emission factor; D route (This indicates the length of the road segment). Simultaneously, the fog layer integrates the pavement degradation model prediction results (calculating the maintenance demand cycle based on the Paris fatigue crack propagation formula), combines it with IoT data from maintenance machinery (such as GNSS trajectories of road rollers and diesel consumption curves), quantifies the carbon emissions from activities such as de-icing agent application and crack repair, and finally sums traffic flow emissions and maintenance emissions to generate the total carbon emissions for a single highway's operating period. Key parameters of the accounting process are recorded using blockchain to ensure audit traceability.

[0139] S334. Utilize cloud layer for micro-traffic simulation, simulate the correlation curve between congestion index and emissions, and calculate the carbon emissions of the entire road network in parallel through road network-level integration.

[0140] Specifically, the process is executed by a high-performance computing cluster in the cloud. First, real-time traffic flow data is imported to drive a vehicle following model (IDM model parameters: maximum acceleration 2.6 m / s², safe distance 2 seconds). This simulates vehicle start-stop and lane-changing behavior under different congestion indices (0-1 continuous values) to generate a cluster of speed-emission relationship curves. The MapReduce parallel framework is used to divide the entire road network into 1km×1km computing units. Each unit deploys an independent simulation instance (a total of thousands of concurrent processes). The emission factor is dynamically adjusted in conjunction with meteorological data (temperature and wind speed fields output by the WRF model).

[0141] Finally, the results of each unit are aggregated through a reduction operation, and a full road network carbon emission hypercube dataset containing spatiotemporal dimensions (longitude, latitude, time slice) and attribute dimensions (vehicle type, fuel type, emission type) is output. It supports multidimensional analysis and visualization, and the computing performance reaches processing millions of vehicle trajectory data per minute.

[0142] S34. Calculate the total carbon emissions during the demolition phase using a hierarchical computing architecture and integrate them throughout the entire life cycle.

[0143] In the description of this invention, the calculation of total carbon emissions during the demolition phase using a hierarchical computing architecture and the integration of the entire lifecycle include:

[0144] S341. Divide the multimodal data of the demolition phase into different demolition emission types and extract the corresponding accounting indicators. The demolition emission types include demolition equipment type and waste treatment type.

[0145] S342. Use the edge layer to identify and record data on waste on highways and usage data of dismantling equipment.

[0146] S343. Utilize the fog layer to generate a unique digital ID for each batch of recycled waste, record the transportation distance, processing technology and reuse projects of the waste, and calculate the total carbon emissions of the demolition during the demolition phase.

[0147] S344. Utilize the cloud layer to perform circular economy offset calculations, integrate and calculate the carbon emission distribution and total carbon emissions throughout the entire life cycle of highways, store the output results of each layer in the cloud, and record the source, processing path and quality label of each data through data lineage tracking.

[0148] S35. Construct a virtual highway model based on digital twins, introduce an incremental online learning mechanism, integrate the computational data of each layer in the hierarchical computing architecture, and realize cross-stage and cross-temporal coupled simulation.

[0149] In the description of this invention, constructing a virtual highway model based on digital twins, introducing an incremental online learning mechanism, and integrating computational data from various layers within a hierarchical computing architecture to achieve cross-stage and cross-temporal coupled simulation includes:

[0150] S351. By combining geographic information systems, building information models, and multiphysics coupling, a virtual highway model based on digital twins is constructed, and the output data of the hierarchical computing architecture is integrated.

[0151] S352. Embed an incremental online learning module in the virtual highway model to regularly update edge and cloud data and adjust and correct the model parameters in the virtual highway model.

[0152] Specifically, an incremental online learning module is embedded in the digital twin model, enabling the model to continuously update itself, adjust parameters based on newly collected data, and achieve real-time prediction and feedback.

[0153] Online learning algorithm design can employ online gradient descent or incremental machine learning algorithms to update model parameters in real time. For example, let the model parameters be θ, and the new data sample be (x... t y t The update formula for online learning can be expressed as:

[0154] ;

[0155] In the formula, L(θ) t ;x t y t ) represents the current loss function (e.g., mean squared error), and α is the learning rate. This formula ensures that the model parameters can be quickly adjusted after receiving new data, reducing prediction error.

[0156] The virtual highway model periodically or in real time feeds new data from the edge and cloud into the online learning module; the online learning module uses the new data to fine-tune the parameters of key sub-models in the digital twin model (such as the carbon emission prediction model and the state monitoring model), without having to start training from scratch, thus achieving rapid adaptive updates.

[0157] S353. Use a virtual highway model to perform coupled simulations of different stages, regions, and time periods throughout the entire life cycle of a highway, simulating the interrelationships between different stages.

[0158] In the description of this invention, a virtual highway model is used to perform coupled simulations of different stages, regions, and time periods throughout the entire life cycle of a highway, simulating the interrelationships between different stages, including:

[0159] S3531. Establish state vectors for the construction phase, operation phase, and demolition phase respectively. An example of the coupling formula is as follows:

[0160] ;

[0161] In the formula, X i (t) represents the state of stage i at time t; f i The evolution function within stage i represents the dynamic changes within that stage; u i (t) represents the external input received in stage i; g ij This represents the coupling effect function of other stages j on stage i, quantifying the transmission effect between different stages.

[0162] S3532. Introduce a multi-scale simulation module into the virtual highway model and use a spatiotemporal coupling simulation algorithm to interactively fuse multimodal data and calculation results from each stage and region.

[0163] S3533. The multi-scale simulation modules at each stage transmit data through a unified interface and adopt a hierarchical simulation strategy to integrate local short-term dynamics with global long-term trends, thereby achieving cross-temporal and spatial simulation.

[0164] S3534. The carbon emission data output by the hierarchical computing architecture is used as the actual accounting value and compared with the carbon emission prediction value of the virtual highway model coupled simulation. The difference between the prediction value and the actual accounting value is quantitatively evaluated, and the difference value is fed back to the incremental online learning module to optimize the parameters of the virtual highway model.

[0165] S354. By constructing a highway knowledge graph, the virtual highway model is adapted for general use.

[0166] In the description of this invention, the generalized adaptation of the virtual highway model by constructing a highway knowledge graph includes:

[0167] S3541. Set up a unified feature space, use a pre-trained model as a feature extractor, freeze the underlying network of the virtual highway model, and build a transfer learning framework.

[0168] Specifically, by constructing a highway knowledge graph, the virtual highway model is made universally adaptable. First, in step S3541, a unified feature space is set and a transfer learning framework is built. This includes standardizing various heterogeneous data (such as traffic flow, material properties, environmental parameters, etc.) involved in the entire life cycle of highways, defining a unified feature vector containing core indicators such as average daily traffic volume, road surface condition index, temperature, and humidity, and using a pre-trained deep neural network model (such as LSTM or graph convolutional network trained based on historical highway project data) as a feature extractor. The weights of its bottom layer network are frozen to retain the general representation ability of basic features, and only the top fully connected layer is fine-tuned to adapt, forming a transferable model architecture.

[0169] In this process, Keras (an open-source artificial neural network library) or PyTorch (an open-source deep learning framework for machine learning and deep learning) is used to implement partial freezing and parameter sharing mechanisms for network layers, and few-shot learning strategies (such as feature alignment based on contrastive learning) are used to improve the model's adaptability to new projects.

[0170] S3542. Construct a highway knowledge graph to store prior knowledge about highways, and when a new project is input, automatically match similar historical cases to initialize the model parameters of the virtual highway model.

[0171] Specifically, constructing a knowledge graph for highways can utilize graph databases (such as Neo4j graph database software) to store entity nodes (including material types, construction techniques, climate zones, etc.) and their relationships in the field of highway engineering. Natural language processing technology can be used to extract structured knowledge from industry standard documents and engineering case libraries to establish a semantic network containing prior knowledge such as material performance parameters, typical structural designs, and regional emission factors.

[0172] When a new project is input, its engineering features (such as geographical coordinates, design speed, and pavement structure layer combination) are first analyzed. A graph embedding algorithm (such as Node2Vec) is used to map the project features to the vector space of the knowledge graph. The cosine similarity with historical case nodes is calculated, and past project cases with topological similarity exceeding a threshold (usually set to 0.85) are automatically matched. The model parameters trained on the corresponding cases (including neural network weights and emission factor library data) are extracted as the initialization parameters for the new model. Simultaneously, a graph attention mechanism (GAT) is used to aggregate the feature information of adjacent nodes, dynamically adjusting the initialization parameters to adapt to regional differences (such as the impact of atmospheric oxygen content on mechanical combustion efficiency in plateau regions). Ultimately, this enables the rapid deployment of virtual simulation models suitable for highway projects with different climate conditions and traffic levels without retraining.

[0173] S4. Draw a carbon footprint heat map based on carbon emission accounting results to show the spatiotemporal evolution process.

[0174] Specifically, carbon emission data is collected from a hierarchical computing architecture (edge-fog-cloud), including carbon emission information from the construction, operation, and demolition phases. Combined with Geographic Information System (GIS) data of highways, carbon emission accounting data is mapped to road grids (e.g., 100m × 100m rasterized areas). Interpolation algorithms (such as Kriging interpolation and inverse distance weighted regression (IDW)) are used to fill in missing data, ensuring the spatial continuity of carbon emissions. The geographically weighted regression (GWR) method is used to calculate the carbon emission intensity of each spatial unit. Visualization tools, such as Matplotlib, Seaborn, Plotly, or Kepler.gl, are used to create a carbon footprint heatmap, displaying carbon emission intensity using a color gradient (blue-green-yellow-red).

[0175] Using GIS spatiotemporal interpolation technology, dynamic charts of the carbon footprint (such as dynamic layers and heatmap animations) are generated. A time slider function allows for the visualization of the carbon footprint over time, analyzing the spatiotemporal distribution changes of carbon emissions at different stages. Furthermore, K-Means clustering or DBSCAN density clustering methods can be used to identify high-value carbon emission areas (hotspots). Combining road topology analysis, key influencing factors of carbon emissions, such as traffic density, construction energy consumption, and transportation distance, are analyzed. Optimization measures are proposed for high-carbon emission areas, such as construction energy consumption control, road network optimization, and intelligent traffic scheduling. By combining real-time monitoring data, dynamic carbon emission management is achieved, gradually reducing the carbon emission intensity of highways throughout their entire lifecycle.

[0176] In summary, by utilizing the above-mentioned technical solutions of this invention, this invention covers the construction, operation, and demolition stages of the entire life cycle of highways, ensuring the comprehensiveness and systematic nature of the accounting results; it employs multiple data sources and collection methods, combined with corresponding accounting methods, to ensure the accuracy and reliability of the accounting results; at the same time, it provides detailed accounting processes and steps, facilitating practical operation and application, and providing a scientific basis for the low-carbon construction and operation of highways.

[0177] By dividing the entire life cycle of highways into construction, operation, and demolition phases, and dynamically defining the accounting boundaries of each phase based on spatiotemporal dimensions, refined carbon emission management is achieved throughout the entire process and at all times. The use of dynamic boundary correction technology accurately captures project progress and environmental changes, ensuring the real-time nature and representativeness of carbon emission data at each phase. This overcomes the limitations of traditional methods, such as single-phase division and incomplete data collection, providing a scientific basis for green construction and low-carbon operation. It not only significantly improves the accuracy and reliability of the accounting results but also provides data support for subsequent carbon reduction strategy optimization and environmental impact assessment.

[0178] By setting carbon emission indicators for all types and collecting multimodal data corresponding to each stage, efficient integration of various types of information is achieved. With the help of a layered computing architecture that integrates edge and cloud, on-site sensing and historical data can be processed in real time, and data fusion algorithms can be used to eliminate the differences caused by a single accounting method. At the same time, the online learning and updating virtual highway model dynamically verifies and self-corrects the calculation results, ensuring that the accounting results at each stage reach a high level in terms of accuracy and comparability. This effectively reduces data noise and errors, and improves the scientificity and adaptability of the whole life cycle carbon emission accounting.

[0179] It should be understood that although the steps in the flowcharts of the accompanying figures are shown sequentially as indicated by the arrows, these steps are not necessarily executed in the order indicated by the arrows. Unless explicitly stated herein, there is no strict order restriction on the execution of these steps, and they can be executed in other orders. Moreover, at least some steps in the flowcharts of the accompanying figures may include multiple sub-steps or multiple stages. These sub-steps or stages are not necessarily completed at the same time, but can be executed at different times, and their execution order is not necessarily sequential, but can be performed alternately or in turn with other steps or at least some of the sub-steps or stages of other steps.

Claims

1. A method for accounting for carbon emissions throughout the life cycle of a highway, characterized by, include: S1. Divide the entire life cycle of the expressway into multiple stages, and dynamically define the accounting boundaries of each stage based on the spatiotemporal dimension. S2. Set up carbon emission accounting indicators for all types of carbon emissions and collect multimodal data for each stage. S3. Build a layered computing architecture based on edge-cloud collaborative technology to calculate carbon emissions at each stage, and use a virtual highway model that can be learned and updated online to verify the carbon emission calculation results. S4. Draw a carbon footprint heat map based on carbon emission accounting results to show the spatiotemporal evolution process; The method of dividing the entire life cycle of a highway into multiple stages and dynamically defining the accounting boundaries of each stage based on spatiotemporal dimensions includes: S11. The entire life cycle of the expressway is divided into the construction phase, the operation phase, and the demolition phase, and the duration of each phase is dynamically adjusted according to the actual progress of the project. S12. Utilize highway engineering information in conjunction with geographic information systems and building information models to define the physical impact range and emission responsibility areas for carbon emissions at each stage. S13. Construct a carbon flow transfer matrix between each stage to quantify the transfer effect of carbon emissions at each stage; The aforementioned hierarchical computing architecture based on edge-cloud collaborative technology is used to calculate carbon emissions at each stage, and the carbon emission calculation results are verified using a virtual highway model that can be learned and updated online. S31. Build a layered computing architecture consisting of an edge layer, a fog layer, and a cloud layer, and deploy it sequentially on the engineering machinery end, the road section management center, and the traffic control platform to realize full life cycle computing of highways; S32. Calculate the total carbon emissions during the construction phase using a hierarchical computing architecture; S33. Calculate the total operational carbon emissions at the road network level during the operational phase using a hierarchical computing architecture. S34. Calculate the total carbon emissions during the demolition phase using a hierarchical computing architecture and integrate them throughout the entire life cycle. S35. Construct a virtual highway model based on digital twins, introduce an incremental online learning mechanism, integrate the computing data of each layer in the hierarchical computing architecture, and realize cross-stage and cross-temporal coupled simulation. The construction of a virtual highway model based on digital twins, the introduction of an incremental online learning mechanism, and the integration of computational data from various layers within a hierarchical computing architecture to achieve coupled simulation across stages and time periods include: S351. Combining geographic information systems, building information models, and multiphysics coupling, a virtual highway model based on digital twins is constructed, and the output data of the hierarchical computing architecture is integrated. S352. Embed an incremental online learning module in the virtual highway model, regularly update edge and cloud data, and adjust and correct the model parameters in the virtual highway model. S353. Use a virtual highway model to perform coupled simulations of different stages, regions, and time periods throughout the entire life cycle of a highway, and simulate the interrelationships between different stages. S354. By constructing a highway knowledge graph, the virtual highway model is adapted for general use.

2. The method according to claim 1, wherein, The dynamic adjustment of the duration of each stage based on the actual progress of the project includes: S111. Obtain the building information model and engineering machinery data of the highway project, divide the highway construction area into several unit grids, and generate a machinery coverage matrix. S112. Compare the actual coverage grid of the highway with the design grid, calculate the real-time progress, and use a long short-term memory network to predict the future construction progress. When the construction progress exceeds the preset progress threshold, switch the highway from the construction phase to the operation phase and mark the construction and operation time information. S113. Obtain the pavement performance index of the expressway. When the pavement performance index is lower than the preset performance threshold, switch the expressway from the operation stage to the demolition stage and mark the demolition time information.

3. The method according to claim 1, wherein, The calculation of total construction carbon emissions during the construction phase using a hierarchical computing architecture includes: S321. Divide the multimodal data of the construction phase into different construction emission types and extract the corresponding accounting indicators. The construction emission types include equipment operation, building material production and transportation. S322. Calculate the single-unit carbon emissions of each construction emission type during the construction phase using the edge layer; S323. Utilize fog layers to unify the mechanical trajectories of construction machinery with the coordinates of the building information model during the construction phase, calibrate the clocks of various equipment, and integrate and calculate the total carbon emissions of the construction phase. S324. Utilize the cloud layer to match the factor library of engineering machinery and building materials, dynamically update the edge layer and fog layer, and monitor the abnormal accounting data of the edge layer and fog layer in real time.

4. The method according to claim 1, wherein, The calculation of total operational carbon emissions at the road network level during the operational phase using a hierarchical computing architecture includes: S331. Divide the multimodal data of the operation phase into different operation emission types and extract the corresponding accounting indicators. The operation emission types include vehicle type and facility maintenance type. S332. Use the edge layer to preprocess the multimodal data and extract the non-stop toll collection data of vehicles, mobile phone signaling data and road monitoring data from various accounting indicators; S333. By integrating electronic non-stop toll collection data and mobile phone signaling data using fog layers, a minute-level traffic flow matrix is ​​generated, traffic flow carbon emissions are calculated, and combined with the calculated road maintenance carbon emissions, the total operating carbon emissions of a single expressway during its operation phase are generated. S334. Utilize cloud layer for micro-traffic simulation, simulate the correlation curve between congestion index and emissions, and calculate the carbon emissions of the entire road network in parallel through road network-level integration.

5. The method according to claim 1, wherein, The calculation of total carbon emissions during the demolition phase using a hierarchical computing architecture, and the integration across the entire lifecycle, includes: S341. Divide the multimodal data of the demolition phase into different demolition emission types and extract the corresponding accounting indicators. The demolition emission types include demolition equipment type and waste treatment type. S342. Use the edge layer to identify and record data on waste on highways and usage data of dismantling equipment; S343. Use fog to generate a unique digital ID for each batch of recycled waste, record the transportation distance, processing technology and reuse projects of the waste, and calculate the total carbon emissions of demolition during the demolition phase. S344. Utilize the cloud layer to perform circular economy offset calculations, integrate and calculate the carbon emission distribution and total carbon emissions throughout the entire life cycle of highways, store the output results of each layer in the cloud, and record the source, processing path and quality label of each data through data lineage tracking.

6. The method for statistical accounting of carbon emissions in the whole life cycle of a highway according to claim 1, characterized in that, The method of using a virtual highway model to perform coupled simulations of different stages, regions, and time periods throughout the entire life cycle of a highway, simulating the interrelationships between different stages, includes: S3531. Establish state vectors for the construction phase, operation phase, and demolition phase respectively; S3532. Introduce a multi-scale simulation module into the virtual highway model and use a spatiotemporal coupling simulation algorithm to interactively fuse multimodal data and calculation results from each stage and region. S3533. The multi-scale simulation modules at each stage transmit data through a unified interface and adopt a hierarchical simulation strategy to integrate local short-term dynamics with global long-term trends, thereby achieving cross-temporal and spatial simulation. S3534. The carbon emission data output by the hierarchical computing architecture is used as the actual accounting value and compared with the carbon emission prediction value of the virtual highway model coupled simulation. The difference between the prediction value and the actual accounting value is quantitatively evaluated, and the difference value is fed back to the incremental online learning module to optimize the parameters of the virtual highway model.

7. The method according to claim 1, wherein, The process of generalizing and adapting the virtual highway model by constructing a highway knowledge graph includes: S3541. Set up a unified feature space, use a pre-trained model as a feature extractor, freeze the underlying network of the virtual highway model, and build a transfer learning framework. S3542. Construct a highway knowledge graph to store prior knowledge about highways, and when a new project is input, automatically match similar historical cases to initialize the model parameters of the virtual highway model.

Images