A comprehensive energy decision support data analysis method and system for smart highways

Abstract

This application provides a comprehensive energy decision support data analysis method and system for smart highways, belonging to the field of smart transportation and energy management technology. The method involves: collecting multi-source data from various energy-consuming nodes, generating node energy data packages, and converging them at a management sub-center. This data is then spatiotemporally correlated with traffic business data to extract energy-traffic correlation features, update the state of the digital twin model, generate a structured dataset, and upload it to a cloud control platform for energy demand and new energy power generation forecasting. Based on the forecast results, real-time system status, external boundary conditions, and carbon emission accounting rules, real-time carbon emission accounting and carbon flow tracking of the highway energy system are performed. A multi-objective optimization decision model is solved to generate a global energy dispatch strategy, which is then distributed to the management sub-center and edge computing nodes to execute energy equipment control commands. This application achieves efficient energy management for highways, reduces energy costs, and improves the level of low-carbon operation.

Description

Technical Field

[0001] This application belongs to the field of intelligent transportation and energy management technology, specifically relating to a comprehensive energy decision support data analysis method and system for intelligent highways. Background Technology

[0002] With the increasing urgency of energy conservation and carbon reduction in the transportation industry, especially in highway systems, the following problems are currently prevalent in highway energy management: First, data is fragmented, forming isolated systems, with energy data from service areas, toll stations, tunnels, charging piles, photovoltaic power stations, and other systems collected independently, lacking effective integration; second, management is extensive, making it difficult to conduct refined and predictive energy scheduling based on real-time data and traffic conditions; third, carbon emission accounting methods rely on traditional approaches, depending on manual reporting and periodic statistics, failing to achieve real-time, accurate measurement and tracking during operation; and fourth, there is a lack of coordination and interaction among energy sources, grids, loads, and storage, resulting in insufficient renewable energy absorption capacity and low overall efficiency.

[0003] While some energy consumption monitoring platforms exist in the current technology, they mostly focus on data display and lack in-depth data analysis and decision support capabilities. Although there is research on charging pile scheduling or energy management systems, it is often limited to single business or local optimization, failing to construct a comprehensive energy decision support solution that integrates data perception, fusion analysis, carbon accounting, and collaborative optimization from the perspective of the entire highway area, all elements, and all businesses. Summary of the Invention

[0004] In a first aspect, embodiments of this application provide a comprehensive energy decision support data analysis method for smart highways, comprising the following steps: S1. Deploy edge computing nodes at each energy-consuming node of the highway, collect local multi-source data from each energy-consuming node, and perform cleaning, formatting and preliminary local calculations to generate standardized node energy data packages. S2. Aggregate the node energy data packets of each energy-consuming node to the management sub-center, and perform spatiotemporal correlation fusion with traffic business data to extract regional energy-traffic correlation features. Then, use the regional energy-traffic correlation features to update the state of the regional digital twin model and generate a regional structured dataset. S3. Upload the regional structured dataset of at least one management sub-center to the cloud control platform, and use the prediction model to predict energy demand and new energy power generation for future periods based on historical and real-time data; S4. Based on the forecast results of future energy demand and new energy power generation, the real-time status of the highway energy system, external boundary conditions, and carbon emission accounting rules, the cloud control platform performs real-time accounting and carbon flow tracking of the carbon emissions of the highway energy system, and solves the optimization decision model with the goal of economic efficiency, low carbon emissions, or high proportion of new energy consumption, and generates a global energy dispatch strategy. S5. The global energy scheduling strategy is distributed to the corresponding management sub-centers and edge computing nodes. The edge computing nodes execute the energy equipment control commands and feed back the status and effect data of the highway energy system after the execution of the energy equipment control commands to the cloud control platform to optimize the iterative updates of the optimization decision model and prediction model.

[0005] Furthermore, the specific steps of step S1 are as follows: S11. Collect local multi-source data from energy-consuming nodes; the local multi-source data includes energy consumption data, new energy power generation data, and environmental monitoring data; Among them, energy consumption data should include at least electricity, current, voltage and power factor; new energy power generation data should include at least photovoltaic power generation and energy storage system charging and discharging status; and environmental monitoring data should include at least temperature, humidity and light intensity. S12. Clean the collected raw local multi-source data, remove missing or abrupt values ​​caused by communication anomalies, and use interpolation to complete the data; S13. Encapsulate the cleaned data according to a predefined structured format to generate structured data containing timestamps, node identifiers, data item identifiers, and data values; S14. Perform preliminary local calculations on structured data using edge computing nodes, at least calculating the real-time total energy consumption of the energy-consuming nodes. and local real-time renewable energy consumption rate The calculation formulas are as follows:

[0006]

[0007] in, Let be the real-time power of the i-th load. Real-time photovoltaic power generation. For real-time discharge power of energy storage, Real-time charging power for energy storage; S15. Encapsulate the structured data and local preliminary calculation results into a standardized node energy data package.

[0008] Furthermore, the specific steps of step S2 are as follows: S21. The management sub-center aggregates node energy data packets uploaded by all energy-consuming nodes within its jurisdiction; S22. Obtain traffic operation data for the corresponding time period and region, wherein the traffic operation data includes at least the cross-sectional traffic flow. and average vehicle speed ; S23. Perform spatiotemporal alignment and fusion of node energy data packets and traffic business data to obtain aligned time-series data, and calculate regional energy-traffic correlation characteristics, wherein the regional energy-traffic correlation characteristics include at least the regional energy consumption intensity per unit traffic flow. Traffic-load correlation coefficient The calculation formulas are as follows:

[0009] in, Let t represent the total energy consumption of the region during time period t. This is determined by the real-time total energy consumption of each energy-consuming node. Summing and integrating yields the result;

[0010] Traffic-load correlation coefficient The Pearson correlation coefficient characterizing the total energy consumption sequence and the traffic flow sequence in the region; S24. Input the regional energy-transportation correlation characteristics, the status of equipment at each energy-consuming node, and environmental monitoring data into the preset regional digital twin model, drive the regional digital twin model to update the virtual equipment status based on the regional energy-transportation correlation characteristics, and visualize the energy flow and traffic flow. S25. Integrate regional energy-transportation correlation features, aligned time-series data, and updated digital twin model equipment status interface data to generate a regional structured dataset.

[0011] Furthermore, the specific steps of step S3 are as follows: S31. Receive at least one regional structured dataset from a management sub-center, wherein the regional structured dataset contains at least a regional historical energy consumption sequence. Historical photovoltaic power generation sequence Historical energy storage charging and discharging sequences and historical traffic flow sequences ; S32. Perform feature engineering on the regional structured dataset to construct an input feature vector X(t) for prediction. The input feature vector X(t) includes at least lagged historical data, date and time features, and regional energy-transportation correlation features. The lagged historical data includes historical energy consumption, historical photovoltaic power generation, historical energy storage charging and discharging power, and historical light intensity sequence. S33. Input the input feature vector X(t) into the pre-trained prediction model, and output the regional energy demand prediction sequence for the next T time periods. and new energy power generation forecast series .

[0012] Furthermore, the pre-trained prediction model in step S33 is a time series prediction model based on a long short-term memory network. The model structure includes an input layer, an output layer, at least one LSTM layer, and at least one fully connected layer. The training process of the prediction model includes: S331. Divide the historically accumulated regional structured dataset into a training set and a validation set according to a preset ratio. The training set is used for model training, and the validation set is used for model performance evaluation. S332. Construct an LSTM prediction model, where the LSTM layer is used to capture the time dependencies in the input feature vector X(t), and the fully connected layer is used to map the output of the LSTM layer to the predicted value; S333. Take the input feature vector X(t) of the training set as the sample feature, add the corresponding real energy demand value or new energy power generation value as the sample label, input it into the LSTM prediction model in the forward propagation mode, calculate the loss between the prediction output and the real label, use the mean square error function as the error function, and use the back propagation algorithm to optimize the network weight of the LSTM prediction model until the loss converges. S334. Use the validation set to evaluate the prediction accuracy of the LSTM prediction model. Once the prediction accuracy reaches a preset threshold, deploy the model to the cloud control platform. When a new regional structured dataset is uploaded from the management sub-center to the cloud control platform, update the parameters of the LSTM prediction model using the new dataset containing feedback effect data through incremental learning or periodic retraining.

[0013] Furthermore, the specific steps of step S4 are as follows: S41. The cloud control platform acquires future energy demand and new energy power generation forecasts, real-time status data of the highway energy system, and external boundary conditions; the real-time status data of the highway energy system includes at least the current state of charge of each energy storage device, the real-time power of key loads, and the real-time output of new energy power generation equipment. S42. Obtain activity level data reflecting cumulative consumption over a period of time, and use a preset emission factor to calculate the real-time carbon emissions and carbon reduction of the highway energy system; the activity level data includes fossil fuel consumption, net purchased electricity, and renewable energy grid connection electricity of highway fixed facilities; S43. Construct an optimization decision-making model with the objectives of minimizing total operating costs, minimizing total carbon emissions, or maximizing the renewable energy absorption rate; The decision variables of the optimization decision model include the planned power of each controllable device in each future time period, and the constraints include the energy dynamic balance constraint starting from the current state of charge of the energy storage device, the upper and lower limits of the device power, and the power balance constraint of the highway energy system. S44. Using the future energy demand and new energy power generation forecasts, the real-time status data of the highway energy system, and external boundary conditions as model parameters, solve the optimization decision model; S45. The optimal solution obtained by solving the optimization decision model is converted into a global energy scheduling strategy that includes specific action instructions for each controlled device.

[0014] Further, in step S42, calculating the real-time carbon emissions of the highway energy system specifically includes: Calculate the first range of direct emissions :

[0015] in, Let be the consumption amount of the i-th type of fossil fuel. , , These are the lower heating value, carbon content per unit calorific value, and carbon oxidation rate of the i-th type of fossil fuel, respectively. Let j be the amount of refrigerant or extinguishing agent emitted. The global warming potential value of the j-th refrigerant or fire extinguishing agent; Calculate indirect emissions in the second range :

[0016] in, This refers to the net purchased electricity volume. This represents the emission factor of the power grid in the region.

[0017] Furthermore, in step S43, the optimization decision model is a mixed-integer linear programming model, and the objective function is to minimize the total cost. :

[0018]

[0019]

[0020] in, The cost of purchasing electricity during period t. To purchase electricity from the grid, The price at which electricity is purchased from the grid; For electricity sales revenue, To sell electricity to the grid, The unit price for selling electricity to the power grid; For operation and maintenance costs; For carbon trading revenue; The power balance constraints of the highway energy system are as follows:

[0021] in, The new energy power generation capacity predicted in step S3, For the energy demand predicted in step S3, and These represent the energy storage charging and discharging power, respectively.

[0022] Furthermore, the specific steps of step S5 are as follows: S51. The cloud control platform parses the global energy scheduling strategy into energy equipment control commands and sends them to the edge computing nodes corresponding to specific devices; S52. Edge computing nodes execute control commands for energy equipment, adjusting the power of charging piles and changing the working mode of energy storage systems; S53. The edge computing node collects the actual operating data of the equipment after the energy equipment control command is executed, and uses it as new local multi-source data to re-execute steps S1 to S4.

[0023] Secondly, embodiments of this application also provide a comprehensive energy decision support data analysis system for smart highways, comprising: Edge computing nodes are deployed at various energy-consuming nodes along the highway to collect local multi-source data from each energy-consuming node, and perform cleaning, formatting, and preliminary local calculations to generate standardized node energy data packages. The management sub-center communicates with at least one edge computing node to aggregate the node energy data packets of each energy-consuming node and perform spatiotemporal correlation fusion with the corresponding traffic business data to extract regional energy-traffic correlation features. It then uses these regional energy-traffic correlation features to update the state of the region's digital twin model and generate a regional structured dataset. The cloud control platform communicates with the management sub-center and is used to receive regional structured datasets from at least one management sub-center. It uses a prediction model to predict energy demand and new energy power generation for future periods based on historical and real-time data. Based on the prediction results, the real-time status of the highway energy system, external boundary conditions, and carbon emission accounting rules, it performs real-time accounting and carbon flow tracking of the highway energy system's carbon emissions. It also solves an optimization decision model with the goals of economy, low carbon emissions, or high proportion of new energy consumption to generate a global energy dispatch strategy. The strategy execution and feedback module is used to distribute the global energy scheduling strategy to the corresponding management sub-center and edge computing nodes. The edge computing nodes execute the energy equipment control commands and feed back the real-time status and effect data of the highway energy system after the command execution to the cloud control platform to optimize the iterative updates of the optimization model and prediction model.

[0024] As can be seen from the above technical solutions, this application has the following advantages: The comprehensive energy decision support data analysis method and system for smart highways provided in this application achieves local data collection and preliminary processing by deploying edge computing nodes at various energy-consuming nodes along the highway. The data is then aggregated to a management sub-center for integrated analysis and uploaded to a cloud control platform for global optimization decision-making, forming a closed-loop energy management system. This system improves the operational efficiency of the highway energy system, achieves rational energy allocation and utilization, and reduces energy consumption costs. An optimization decision-making model is constructed with economic efficiency, low carbon emissions, or high-proportion renewable energy consumption as objectives. Real-time calculation of carbon emissions from the highway energy system and tracking of carbon flows reduce carbon emissions during highway operation, promoting the highway industry towards low-carbon and green development. Deep mining and analysis of multi-source data through big data analysis, digital twins, and machine learning extracts valuable energy-traffic correlation features, providing a basis for energy dispatching strategies, improving the accuracy and reliability of decision-making, and enhancing the intelligence level of the highway energy system. By analyzing the correlation characteristics between traffic business data and energy data, deep integration and collaborative optimization of traffic flow and energy flow are achieved, which not only alleviates traffic congestion but also improves energy utilization efficiency and enhances the overall operational benefits of the highway. Attached Figure Description

[0025] To more clearly illustrate the technical solution of this application, the accompanying drawings used in the description will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0026] Figure 1 This is a flowchart illustrating the integrated energy decision support data analysis method for smart highways according to the present invention.

[0027] Figure 2 This is a schematic diagram of the integrated energy decision support data analysis system for smart highways according to the present invention. Detailed Implementation

[0028] The various embodiments of this disclosure will be described more fully in the following detailed description of the specific steps of the integrated energy decision support data analysis method for smart highways. This disclosure may have various embodiments, and adjustments and changes may be made therein. However, it should be understood that there is no intention to limit the various embodiments of this disclosure to the specific embodiments disclosed herein, but rather this disclosure should be understood to cover all adjustments, equivalents, and / or alternatives falling within the spirit and scope of the various embodiments of this disclosure.

[0029] This embodiment provides a comprehensive energy decision support data analysis method for smart highways, enabling efficient energy management of highways, reducing energy costs, and improving low-carbon operation levels.

[0030] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0031] Please see Figure 1 The diagram shows a flowchart of a comprehensive energy decision support data analysis method for smart highways in a specific embodiment. The method includes the following steps: S1. Deploy edge computing nodes at each energy-consuming node of the highway, collect local multi-source data from each energy-consuming node, and perform cleaning, formatting and preliminary local calculations to generate standardized node energy data packages. It should be noted that by deploying edge computing nodes at various energy-consuming nodes along the highway, comprehensive local multi-source data from each node can be collected, providing a rich and comprehensive data source for energy analysis and decision-making, thus avoiding decision-making errors caused by data gaps. The collected data is cleaned, formatted, and subjected to preliminary local calculations to generate standardized node energy data packages. This effectively removes noise and outliers from the data, improving data quality. Simultaneously, the data is converted into a unified format, facilitating subsequent data aggregation and fusion processing. Furthermore, preliminary local calculations can quickly extract key energy indicators from energy-consuming nodes, supporting real-time energy monitoring and preliminary analysis, and improving the efficiency and real-time performance of data processing. S2. Aggregate the node energy data packets of each energy-consuming node to the management sub-center, and perform spatiotemporal correlation fusion with traffic business data to extract regional energy-traffic correlation features. Then, use the regional energy-traffic correlation features to update the state of the regional digital twin model and generate a regional structured dataset. It should be noted that aggregating the node energy data packets from each energy-consuming node to the management sub-center and performing spatiotemporal correlation and fusion with traffic business data can deeply explore the intrinsic relationship between energy data and traffic data, extract regional energy-traffic correlation characteristics, and thus provide a more comprehensive understanding of the mutual influence between highway energy consumption and traffic operation, providing an important basis for the coordinated optimization of energy scheduling and traffic management. The regional energy-traffic correlation characteristics are used to update the state of the regional digital twin model, generating a regional structured dataset. Through the digital twin, the energy flow and traffic flow status within the region, as well as the virtual operation of equipment, can be presented in real time and intuitively, providing managers with a virtual decision-making environment consistent with reality. Simultaneously, timely updates to the digital twin model state ensure that the model always reflects the latest operational status, providing an accurate model for analysis and decision-making. S3. Upload the regional structured dataset of at least one management sub-center to the cloud control platform, and use the prediction model to predict energy demand and new energy power generation for future periods based on historical and real-time data; It should be noted that at least one regional structured dataset from a management sub-center is uploaded to the cloud control platform. Based on historical and real-time data, energy demand and renewable energy generation forecasts for future periods are made. The cloud control platform aggregates data from multiple regions, enabling analysis of the energy system's operational patterns from a macro perspective. Using big data analytics and advanced forecasting models, such as time series forecasting models based on Long Short-Term Memory (LSTM) networks or Transformer architectures, it can accurately predict future energy demand and renewable energy generation, providing a reliable basis for formulating energy dispatch strategies and improving the energy system's responsiveness. Uploading regional structured datasets to the cloud control platform enables the integration and sharing of data from different regions. This allows the cloud control platform to manage and coordinate the energy system of the entire highway network in a unified manner, achieve globally optimal energy dispatch strategies, and improve the overall operating efficiency of the highway energy system. S4. Based on the forecast results of future energy demand and new energy power generation, the real-time status of the highway energy system, external boundary conditions, and carbon emission accounting rules, the cloud control platform performs real-time accounting and carbon flow tracking of the carbon emissions of the highway energy system, and solves the optimization decision model with the goal of economic efficiency, low carbon emissions, or high proportion of new energy consumption, and generates a global energy dispatch strategy. It should be noted that this step comprehensively considers multiple factors and objectives. Through reasonable model construction and solution methods, the optimal energy dispatch strategy can be obtained, realizing multi-objective optimization of the highway energy system, making energy dispatch decisions reasonable, and taking into account economic benefits, environmental benefits, and the efficiency of new energy utilization. Real-time accounting and carbon flow tracking of the carbon emissions of the highway energy system can accurately grasp the carbon emission situation of the highway energy system, providing data support for carbon emission management, and also providing accurate carbon emission data basis for highway operating companies to participate in carbon trading, etc. S5. The global energy scheduling strategy is distributed to the corresponding management sub-center and edge computing nodes. The edge computing nodes execute the energy equipment control commands and feed back the real-time status and effect data of the highway energy system after the energy equipment control commands are executed to the cloud control platform to optimize the iterative update of the optimization decision model and prediction model. It should be noted that the timely strategy execution and feedback mechanism in this step ensures the effective implementation of energy dispatch strategies. Furthermore, the cloud control platform can promptly understand the effectiveness of strategy execution based on feedback data, evaluate and adjust the actual operating conditions, thereby improving the system's response speed and adaptability. The feedback system status and effect data are used to optimize the iterative updates of the model, enabling the model to continuously learn and improve based on actual operating conditions, gradually enhancing the model's accuracy and adaptability.

[0032] This embodiment deploys edge computing nodes at various energy-consuming nodes along the highway to collect multi-source data and perform cleaning, fusion, and analysis, thereby achieving accurate prediction and optimized scheduling of energy demand, reducing energy costs, decreasing carbon emissions, and improving the synergistic benefits of transportation and energy.

[0033] Furthermore, as a refinement and extension of the specific implementation methods of the above embodiments, in order to fully illustrate the specific implementation process of this embodiment, another comprehensive energy decision support data analysis method for smart highways is provided, which includes the following steps: S1. Deploy edge computing nodes at each energy-consuming node along the highway, collect local multi-source data from each node, and perform cleaning, formatting, and preliminary local calculations to generate standardized node energy data packages; the specific steps of step S1 are as follows: S11. Collect local multi-source data from energy-consuming nodes; the local multi-source data includes energy consumption data, new energy power generation data, and environmental monitoring data; Among them, energy consumption data should include at least electricity, current, voltage and power factor; new energy power generation data should include at least photovoltaic power generation and energy storage system charging and discharging status; and environmental monitoring data should include at least temperature, humidity and light intensity. S12. Clean the collected raw local multi-source data, remove missing or abrupt values ​​caused by communication anomalies, and use interpolation to complete the data; S13. Encapsulate the cleaned data according to a predefined structured format to generate structured data containing timestamps, node identifiers, data item identifiers, and data values; S14. Perform preliminary local calculations on structured data using edge computing nodes, at least calculating the real-time total energy consumption of the energy-consuming nodes. and local real-time renewable energy consumption rate The calculation formulas are as follows:

[0034]

[0035] in, Let be the real-time power of the i-th load. Real-time photovoltaic power generation. For real-time discharge power of energy storage, Real-time charging power for energy storage; S15. Encapsulate the structured data and local preliminary calculation results into a standardized node energy data package; For example, taking a service area charging station on a highway section as an energy-consuming node, after deploying an edge computing node, the energy consumption data of the node within 1 hour is collected: cumulative electricity consumption 800kWh, average current 120A, average voltage 380V, power factor 0.92; new energy power generation data: peak photovoltaic power generation 35kW, energy storage system charging state for 40 minutes (charging power 20kW), discharging state for 20 minutes (discharging power 15kW); environmental monitoring data: average temperature 28℃, average humidity 65%, average light intensity 8000lux. During data cleaning, a current jump of 500A was observed in a 5-minute period (far exceeding the normal range). This was identified as an abnormal communication jump and removed. Using linear interpolation, the current data for that period was completed to 120A based on the adjacent 5-minute data points of 118A and 122A. An example of the formatted structured data is as follows: timestamp "2024-08-01 10:00:00", node identifier "FW-QY-003", data item identifier "ELEC-POWER". The data value is "38.5kW". In the initial local calculation, the load of this node includes charging piles (3, with real-time power of 12kW, 15kW and 10kW respectively), lighting equipment (2kW) and air conditioning equipment (8kW). The real-time total energy consumption is calculated according to the formula as 12+15+10+2+8=47kW. The local real-time renewable energy absorption rate is calculated as (35+15) / 47≈1.06, that is, 106%, which indicates that the renewable energy power generation can fully cover the load demand during this period and there is a surplus. S2. Aggregate the node energy data packets of each energy-consuming node to the management sub-center, and perform spatiotemporal correlation fusion with traffic business data to extract regional energy-traffic correlation features. Then, use the regional energy-traffic correlation features to update the state of the regional digital twin model and generate a regional structured dataset. The specific steps of step S2 are as follows: S21. The management sub-center aggregates node energy data packets uploaded by all energy-consuming nodes within its jurisdiction; S22. Obtain traffic operation data for the corresponding time period and region, wherein the traffic operation data includes at least the cross-sectional traffic flow. and average vehicle speed ; S23. Perform spatiotemporal alignment and fusion of node energy data packets and traffic business data to obtain aligned time-series data, and calculate regional energy-traffic correlation characteristics, wherein the regional energy-traffic correlation characteristics include at least the regional energy consumption intensity per unit traffic flow. Traffic-load correlation coefficient The calculation formulas are as follows:

[0036] in, Let t represent the total energy consumption of the region during time period t. This is determined by the real-time total energy consumption of each energy-consuming node. Summing and integrating yields the result;

[0037] Traffic-load correlation coefficient The Pearson correlation coefficient characterizing the total energy consumption sequence and the traffic flow sequence in the region; It should be noted that spatiotemporal alignment and fusion specifically involve: using a preset time granularity Δt (e.g., 15 minutes) as a window, resampling and aligning time-series data from different data sources, and using the physical jurisdiction of the management sub-center as the spatial boundary to spatially aggregate energy-consuming node data; S24. Input the regional energy-transportation correlation characteristics, the status of equipment at each energy-consuming node, and environmental monitoring data into the preset regional digital twin model, drive the regional digital twin model to update the virtual equipment status based on the regional energy-transportation correlation characteristics, and visualize the energy flow and traffic flow. S25. Integrate regional energy-transportation correlation characteristics, aligned time-series data, and updated digital twin model equipment status interface data to generate a regional structured dataset; For example, a management sub-center manages a 30km section of highway, covering 2 service areas, 1 toll station, and 3 tunnel energy consumption nodes. After aggregating the node energy data packets from each node between 10:00 and 11:00 on the same day, it obtains the traffic business data for the area during that time period: cross-sectional traffic flow of 1200 vehicles / hour and average vehicle speed of 85km / h. For spatiotemporal alignment, the energy consumption data of each energy consumption node is matched and fused with the traffic flow data of the corresponding time period at a 15-minute time granularity. When calculating the energy consumption intensity per unit traffic flow in the area, the total energy consumption of each energy consumption node in real time is first summed and integrated to obtain the total energy consumption of the area in time period t (1 hour) as 2800kWh. According to the formula, the energy consumption intensity per unit traffic flow is calculated as 2800 / 1200≈2.33kWh / vehicle. When calculating the traffic-load correlation coefficient, the correlation coefficient between the total energy consumption sequence and the traffic flow sequence in the area during that time period is calculated using the Pearson correlation coefficient formula, which shows that the correlation coefficient is 0.82, indicating that the two are strongly positively correlated. These related features, the status of each node device (such as the normal operation of 3 toll collection devices at the toll station and the power of the tunnel lighting equipment of 50kW), and environmental monitoring data are input into the regional digital twin model. The photovoltaic power generation and energy storage battery charge status virtually presented in the model are synchronized with the actual data in real time. The energy flow shows the transmission path from the photovoltaic power station to the charging pile and lighting equipment, while the traffic flow dynamically displays the driving status of vehicles in the road segment. After the model status is updated, a regional structured dataset containing all the above data is generated. S3. Upload the regional structured dataset of at least one management sub-center to the cloud control platform, and use a predictive model to predict energy demand and new energy power generation for future periods based on historical and real-time data; the specific steps of step S3 are as follows: S31. Receive at least one regional structured dataset from a management sub-center, wherein the regional structured dataset contains at least a regional historical energy consumption sequence. Historical photovoltaic power generation sequence Historical energy storage charging and discharging sequences and historical traffic flow sequences ; S32. Perform feature engineering on the regional structured dataset to construct an input feature vector X(t) for prediction. The input feature vector X(t) includes at least lagged historical data, date and time features, and regional energy-transportation correlation features. The lagged historical data includes historical energy consumption, historical photovoltaic power generation, historical energy storage charging and discharging power, and historical light intensity sequence. S33. Input the input feature vector X(t) into the pre-trained prediction model, and output the regional energy demand prediction sequence for the next T time periods (e.g., the next 24 hours, with Δt as the interval). and new energy power generation forecast series The prediction model is a time series prediction model based on a Long Short-Term Memory (LSTM) network or a Transformer architecture; the pre-trained prediction model in step S33 is a time series prediction model based on a Long Short-Term Memory network, and the model structure includes an input layer, an output layer, at least one LSTM layer, and at least one fully connected layer; the training process of the prediction model includes: S331. Divide the historically accumulated regional structured dataset into a training set and a validation set according to a preset ratio. The training set is used for model training, and the validation set is used for model performance evaluation. S332. Construct an LSTM prediction model, where the LSTM layer is used to capture the time dependencies in the input feature vector X(t), and the fully connected layer is used to map the output of the LSTM layer to the predicted value; S333. Take the input feature vector X(t) of the training set as the sample feature, add the corresponding real energy demand value or new energy power generation value as the sample label, input it into the LSTM prediction model in the forward propagation mode, calculate the loss between the prediction output and the real label, use the mean square error function as the error function, and use the back propagation algorithm to optimize the network weight of the LSTM prediction model until the loss converges. S334. Use the validation set to evaluate the prediction accuracy of the LSTM prediction model. When the prediction accuracy reaches the preset threshold, deploy the model to the cloud control platform. When a new regional structured dataset is uploaded from the management sub-center to the cloud control platform, update the parameters of the LSTM prediction model using the new dataset containing feedback effect data in the form of incremental learning or periodic retraining. For example, the cloud control platform receives regional structured datasets from three management sub-centers, which include the regional historical energy consumption sequence, historical new energy power generation sequence, and historical traffic flow sequence for the past year. In the feature engineering phase, an input feature vector X(t) is constructed. The lagged historical data consists of energy consumption, power generation, and traffic flow data from the same time period over the past 7 days. Date and time features include whether it is a weekday (yes), whether it is a holiday (no), and the time period (14:00-15:00). The regional energy-traffic correlation features are energy consumption intensity per unit traffic flow of 2.2 kW / h / vehicle and a traffic-load correlation coefficient of 0.78. This input feature vector is then fed into a pre-trained LSTM prediction model. The model outputs a regional energy demand prediction sequence for the next 24 time periods (each 15 minutes). The predicted energy demand for the next 14:00-14:15 is 320 kW, and for the next 14:15-14:30 it is 335 kW. In the new energy power generation prediction sequence, the corresponding photovoltaic power generation predictions for the time periods are 280 kW and 295 kW, respectively. Subsequently, after a new regional structured dataset is uploaded, it is found that the actual energy consumption for a certain time period for three consecutive days is 15% higher than the predicted value. Through incremental learning, the model parameters are updated with a new dataset containing this feedback data, improving the subsequent prediction accuracy to 92%. S4. Based on future energy demand and renewable energy generation forecasts, real-time system status, external boundary conditions, and preset carbon emission accounting rules, the cloud control platform performs real-time calculation and carbon flow tracking of carbon emissions from the highway energy system. It then solves an optimization decision model aimed at economic efficiency, low carbon emissions, or high-proportion renewable energy consumption to generate a global energy dispatch strategy. The specific steps of step S4 are as follows: S41. The cloud control platform acquires future energy demand and new energy power generation forecasts, real-time status data of the highway energy system, and external boundary conditions; the real-time status data of the highway energy system includes at least the current state of charge of each energy storage device, the real-time power of key loads, and the real-time output of new energy power generation equipment. S42. Obtain activity level data reflecting cumulative consumption over a period of time, and use a preset emission factor to calculate the real-time carbon emissions and carbon reductions of the highway energy system; the activity level data includes fossil fuel consumption, net purchased electricity, and renewable energy grid connection electricity of highway fixed facilities (e.g., service areas, toll stations, tunnels); in step S42, calculating the real-time carbon emissions of the highway energy system specifically includes: Calculate the first range of direct emissions :

[0038] in, Let be the consumption amount of the i-th type of fossil fuel. , , These are the lower heating value, carbon content per unit calorific value, and carbon oxidation rate of the i-th type of fossil fuel, respectively. Let j be the amount of refrigerant or extinguishing agent emitted. The global warming potential value of the j-th refrigerant or fire extinguishing agent; Calculate indirect emissions in the second range :

[0039] in, This refers to the net purchased electricity volume. The emission factor of the power grid in the region; S43. Construct an optimization decision-making model with the objectives of minimizing total operating costs, minimizing total carbon emissions, or maximizing the renewable energy absorption rate; The decision variables of the optimization decision model include the planned power of each controllable device in each future time period, and the constraints include the energy dynamic balance constraint starting from the current state of charge of the energy storage device, the upper and lower limits of the device power, and the power balance constraint of the highway energy system; in step S43, the optimization decision model is a mixed integer linear programming model, and the objective function is to minimize the total cost. :

[0040]

[0041]

[0042] in, The cost of purchasing electricity during period t. To purchase electricity from the grid, The price at which electricity is purchased from the grid; For electricity sales revenue, To sell electricity to the grid, The unit price for selling electricity to the power grid; For operation and maintenance costs; For carbon trading revenue; The power balance constraints of the highway energy system are as follows:

[0043] in, The new energy power generation capacity predicted in step S3, For the energy demand predicted in step S3, and S44 represents the energy storage charging and discharging power, respectively. The optimization decision model is solved using the future energy demand and new energy power generation forecast results, the real-time status data of the highway energy system, and the external boundary conditions as model parameters. S45. The optimal solution obtained by solving the optimization decision model is converted into a global energy scheduling strategy that includes specific action instructions for each controlled device; For example, the cloud control platform obtains future energy demand forecasts (e.g., an average demand of 350kW for each period in the next 2 hours) and new energy power generation forecasts (an average photovoltaic power generation of 290kW); real-time system status data: the current state of charge of the two energy storage devices corresponds to energy values ​​of 80kWh and 95kWh respectively, the real-time power of key loads (charging piles, tunnel lighting) is 340kW, and the real-time output of photovoltaic power is 285kW; external boundary conditions: the grid purchase price of electricity is 1.0 yuan / kWh, and the unit price of electricity sales is 0. The carbon trading price is 50 yuan / ton CO2, priced at 5 yuan / kWh. For carbon emission accounting, the activity level data is as follows: fossil fuel (diesel) consumption is 50 kg (used for emergency generators), net purchased electricity is 1200 kWh, and renewable energy electricity is 300 kWh. Calculating the first-scope direct emissions: diesel has a lower heating value of 42705 kJ / kg, a carbon content per unit calorific value of 20.2 tC / TJ, and a carbon oxidation rate of 0.98. Therefore, the carbon emissions from diesel combustion are 50 × 42705 × 20.2 × 0.98 × 10⁻⁶. -9 ≈0.042 tons of CO2; no refrigerant escape, therefore the total direct emissions in the first area are 0.042 tons of CO2; indirect emissions in the second area: the regional power grid emission factor is 0.6 tCO2 / MWh, therefore the indirect emissions are 1200 × 10 - ³ × 0.6 = 0.72 tons of CO2. The total real-time carbon emissions of the highway energy system are 0.042 + 0.72 = 0.762 tons of CO2. The optimization decision model aims to minimize the total cost. The decision variables include the charging and discharging power of energy storage in each future time period and the power purchased and sold by the grid. The constraints are that the upper limit of energy storage charging power is 30kW and the upper limit of discharging power is 25kW. The system power balance constraint is "power purchased + photovoltaic power generation + energy storage discharging power = load demand power + power sold". The optimal solution is obtained after solving: In the next hour, the energy storage device charges at 20kW for 30 minutes, and the state of charge increases to 90kWh; the power purchased by the grid is maintained at 50kW, and there is no power sale. Based on this, a global energy dispatch strategy is generated, which clarifies the power plan of each controllable device in each time period. S5. The global energy scheduling strategy is distributed to the corresponding management sub-centers and edge computing nodes. The edge computing nodes execute the energy equipment control commands and feed back the system status and effect data after the energy equipment control commands are executed to the cloud control platform to optimize the iterative update of the model. The specific steps of step S5 are as follows: S51. The cloud control platform parses the global energy scheduling strategy into energy equipment control commands and sends them to the edge computing nodes corresponding to specific devices; S52. Edge computing nodes execute control commands for energy equipment, adjusting the power of charging piles and changing the working mode of energy storage systems; S53. The edge computing node collects the actual operating data of the equipment after the energy equipment control command is executed, and uses it as new local multi-source data to re-execute steps S1 to S4; For example, the cloud control platform parses the above-mentioned global energy dispatch strategy into specific energy equipment control commands, such as "Energy storage device FW-ESS-001: Charge at 20kW power from 14:00 to 14:30" and "Grid interface FW-GRID-002: Maintain power purchase at 50kW from 14:00 to 15:00", and sends them to the corresponding management sub-center and edge computing nodes; after receiving the commands, the edge computing nodes control the energy storage device to switch to charging mode and adjust the charging power to 20kW; adjust the power purchase interface of the grid to stabilize at 50kW; after the commands are executed, the edge computing nodes collect the actual operating data of the equipment: energy storage device The actual charging power is 19.8kW, and 9.9kWh of electricity is charged within 30 minutes, increasing the state of charge from 80kWh to 89.9kWh; the actual power purchased by the grid is 50.2kW, and 50.2kWh of electricity is purchased within 1 hour; the actual power consumed by the load is 342kW, and the actual output of photovoltaic power generation is 288kW. The renewable energy absorption rate is (288+19.8) / 342≈0.90, i.e., 90%; these system status and effect data are fed back to the cloud control platform, which uses this data to iteratively update the LSTM prediction model and the optimization decision model, so that the deviation between the subsequent predicted energy demand and the actual demand is reduced to within 5%.

[0044] It should be understood that the sequence number of each step in the above embodiments does not imply the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present invention.

[0045] like Figure 2 As shown, the following are embodiments of the integrated energy decision support data analysis system for smart highways provided in this disclosure. This system and the integrated energy decision support data analysis method for smart highways in the above embodiments belong to the same inventive concept. For details not described in detail in the embodiments of the integrated energy decision support data analysis system for smart highways, please refer to the embodiments of the integrated energy decision support data analysis method for smart highways described above.

[0046] The system includes: Edge computing nodes are deployed at various energy-consuming nodes along the highway to collect local multi-source data from each energy-consuming node, and perform cleaning, formatting, and preliminary local calculations to generate standardized node energy data packages. The management sub-center communicates with at least one edge computing node to aggregate the node energy data packets of each energy-consuming node and perform spatiotemporal correlation fusion with the corresponding traffic business data to extract regional energy-traffic correlation features. It then uses these regional energy-traffic correlation features to update the state of the region's digital twin model and generate a regional structured dataset. The cloud control platform communicates with the management sub-center and is used to receive regional structured datasets from at least one management sub-center. It uses a prediction model to predict energy demand and new energy power generation for future periods based on historical and real-time data. Based on the prediction results, the real-time status of the highway energy system, external boundary conditions, and carbon emission accounting rules, it performs real-time accounting and carbon flow tracking of the highway energy system's carbon emissions. It also solves an optimization decision model with the goals of economy, low carbon emissions, or high proportion of new energy consumption to generate a global energy dispatch strategy. The strategy execution and feedback module is used to distribute the global energy scheduling strategy to the corresponding management sub-center and edge computing nodes. The edge computing nodes execute the energy equipment control commands and feed back the real-time status and effect data of the highway energy system after the command execution to the cloud control platform to optimize the iterative update of the optimization decision model and prediction model.

[0047] This embodiment achieves efficient energy utilization and low-carbon operation, and improves the synergistic benefits of transportation and energy through the interactive collaboration of edge computing node modules, management sub-center platform modules, cloud control platform modules, and strategy execution and feedback modules.

[0048] The above description of the disclosed embodiments enables those skilled in the art to make or use the invention. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the invention. Therefore, the invention is not to be limited to the embodiments shown herein, but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A comprehensive energy decision support data analysis method for smart highways, characterized in that, Includes the following steps: S1. Deploy edge computing nodes at each energy-consuming node of the highway, collect local multi-source data from each energy-consuming node, and perform cleaning, formatting and preliminary local calculations to generate standardized node energy data packages. S2. Aggregate the node energy data packets of each energy-consuming node to the management sub-center, and perform spatiotemporal correlation fusion with traffic business data to extract regional energy-traffic correlation features. Then, use the regional energy-traffic correlation features to update the state of the regional digital twin model and generate a regional structured dataset. S3. Upload the regional structured dataset of at least one management sub-center to the cloud control platform, and use the prediction model to predict energy demand and new energy power generation for future periods based on historical and real-time data; S4. Based on the forecast results of future energy demand and new energy power generation, the real-time status of the highway energy system, external boundary conditions, and carbon emission accounting rules, the cloud control platform performs real-time accounting and carbon flow tracking of the carbon emissions of the highway energy system, and solves the optimization decision model with the goal of economic efficiency, low carbon emissions, or high proportion of new energy consumption, and generates a global energy dispatch strategy. S5. The global energy scheduling strategy is distributed to the corresponding management sub-centers and edge computing nodes. The edge computing nodes execute the energy equipment control commands and feed back the real-time status and effect data of the highway energy system after the execution of the energy equipment control commands to the cloud control platform to optimize the iterative updates of the optimization decision model and prediction model.

2. The integrated energy decision support data analysis method for smart highways according to claim 1, characterized in that, The specific steps of step S1 are as follows: S11. Collect local multi-source data from energy-consuming nodes; the local multi-source data includes energy consumption data, new energy power generation data, and environmental monitoring data; Among them, energy consumption data should include at least electricity, current, voltage and power factor; new energy power generation data should include at least photovoltaic power generation and energy storage system charging and discharging status; and environmental monitoring data should include at least temperature, humidity and light intensity. S12. Clean the collected raw local multi-source data, remove missing or abrupt values ​​caused by communication anomalies, and use interpolation to complete the data; S13. Encapsulate the cleaned data according to a predefined structured format to generate structured data containing timestamps, node identifiers, data item identifiers, and data values; S14. Perform preliminary local calculations on structured data using edge computing nodes, at least calculating the real-time total energy consumption of the energy-consuming nodes. and local real-time renewable energy consumption rate The calculation formulas are as follows: in, Let be the real-time power of the i-th load. Real-time photovoltaic power generation. For real-time discharge power of energy storage, Real-time charging power for energy storage; S15. Encapsulate the structured data and local preliminary calculation results into a standardized node energy data package.

3. The integrated energy decision support data analysis method for smart highways according to claim 1, characterized in that, The specific steps of step S2 are as follows: S21. The management sub-center aggregates node energy data packets uploaded by all energy-consuming nodes within its jurisdiction; S22. Obtain traffic operation data for the corresponding time period and region, wherein the traffic operation data includes at least the cross-sectional traffic flow. and average vehicle speed ; S23. Perform spatiotemporal alignment and fusion of node energy data packets and traffic business data to obtain aligned time-series data, and calculate regional energy-traffic correlation characteristics, wherein the regional energy-traffic correlation characteristics include at least the regional energy consumption intensity per unit traffic flow. Traffic-load correlation coefficient The calculation formulas are as follows: in, Let t represent the total energy consumption of the region during time period t. This is determined by the real-time total energy consumption of each energy-consuming node. Summing and integrating yields the result; Traffic-load correlation coefficient The Pearson correlation coefficient characterizing the total energy consumption sequence and the traffic flow sequence in the region; S24. Input the regional energy-transportation correlation characteristics, the status of equipment at each energy-consuming node, and environmental monitoring data into the preset regional digital twin model, drive the regional digital twin model to update the virtual equipment status based on the regional energy-transportation correlation characteristics, and visualize the energy flow and traffic flow. S25. Integrate regional energy-transportation correlation features, aligned time-series data, and updated digital twin model equipment status interface data to generate a regional structured dataset.

4. The integrated energy decision support data analysis method for smart highways according to claim 1, characterized in that, The specific steps of step S3 are as follows: S31. Receive at least one regional structured dataset from a management sub-center, wherein the regional structured dataset contains at least a regional historical energy consumption sequence. Historical photovoltaic power generation sequence Historical energy storage charging and discharging sequences and historical traffic flow sequences ; S32. Perform feature engineering on the regional structured dataset to construct an input feature vector X(t) for prediction. The input feature vector X(t) includes at least lagged historical data, date and time features, and regional energy-transportation correlation features. The lagged historical data includes historical energy consumption, historical photovoltaic power generation, historical energy storage charging and discharging power, and historical light intensity sequence. S33. Input the input feature vector X(t) into the pre-trained prediction model, and output the regional energy demand prediction sequence for the next T time periods. and new energy power generation forecast series .

5. The integrated energy decision support data analysis method for smart highways according to claim 4, characterized in that, The prediction model pre-trained in step S33 is a time series prediction model based on a long short-term memory network. The model structure includes an input layer, an output layer, at least one LSTM layer, and at least one fully connected layer. The training process for the prediction model includes: S331. Divide the historically accumulated regional structured dataset into a training set and a validation set according to a preset ratio. The training set is used for model training, and the validation set is used for model performance evaluation. S332. Construct an LSTM prediction model, where the LSTM layer is used to capture the time dependencies in the input feature vector X(t), and the fully connected layer is used to map the output of the LSTM layer to the predicted value; S333. Take the input feature vector X(t) of the training set as the sample feature, add the corresponding real energy demand value or new energy power generation value as the sample label, input it into the LSTM prediction model in the forward propagation mode, calculate the loss between the prediction output and the real label, use the mean square error function as the error function, and use the back propagation algorithm to optimize the network weight of the LSTM prediction model until the loss converges. S334. Use the validation set to evaluate the prediction accuracy of the LSTM prediction model. Once the prediction accuracy reaches a preset threshold, deploy the model to the cloud control platform. When a new regional structured dataset is uploaded from the management sub-center to the cloud control platform, update the parameters of the LSTM prediction model using the new dataset containing feedback effect data through incremental learning or periodic retraining.

6. The integrated energy decision support data analysis method for smart highways according to claim 1, characterized in that, The specific steps of step S4 are as follows: S41. The cloud control platform acquires future energy demand and new energy power generation forecasts, real-time status data of the highway energy system, and external boundary conditions; the real-time status data of the highway energy system includes at least the current state of charge of each energy storage device, the real-time power of key loads, and the real-time output of new energy power generation equipment. S42. Obtain activity level data reflecting cumulative consumption over a period of time, and use a preset emission factor to calculate the real-time carbon emissions and carbon reduction of the highway energy system; the activity level data includes fossil fuel consumption, net purchased electricity, and renewable energy grid connection electricity of highway fixed facilities; S43. Construct an optimization decision-making model with the objectives of minimizing total operating costs, minimizing total carbon emissions, or maximizing the renewable energy absorption rate; The decision variables of the optimization decision model include the planned power of each controllable device in each future time period, and the constraints include the energy dynamic balance constraint starting from the current state of charge of the energy storage device, the upper and lower limits of the device power, and the power balance constraint of the highway energy system. S44. Using the future energy demand and new energy power generation forecasts, the real-time status data of the highway energy system, and external boundary conditions as model parameters, solve the optimization decision model; S45. The optimal solution obtained by solving the optimization decision model is converted into a global energy scheduling strategy that includes specific action instructions for each controlled device.

7. The integrated energy decision support data analysis method for smart highways according to claim 6, characterized in that, In step S42, calculating the real-time carbon emissions of the highway energy system specifically includes: Calculate the first range of direct emissions : in, Let be the consumption amount of the i-th type of fossil fuel. , , These are the lower heating value, carbon content per unit calorific value, and carbon oxidation rate of the i-th type of fossil fuel, respectively. Let j be the amount of refrigerant or extinguishing agent emitted. The global warming potential value of the j-th refrigerant or fire extinguishing agent; Calculate indirect emissions in the second range : in, This refers to the net purchased electricity volume. This represents the emission factor of the power grid in the region.

8. The integrated energy decision support data analysis method for smart highways according to claim 6, characterized in that, In step S43, the optimization decision model is a mixed-integer linear programming model, and the objective function is to minimize the total cost. : in, The cost of purchasing electricity during period t. To purchase electricity from the grid, The price at which electricity is purchased from the grid; For electricity sales revenue, To sell electricity to the grid, The unit price for selling electricity to the power grid; For operation and maintenance costs; For carbon trading revenue; The power balance constraints of the highway energy system are as follows: in, The new energy power generation capacity predicted in step S3, For the energy demand predicted in step S3, and These represent the energy storage charging and discharging power, respectively.

9. The integrated energy decision support data analysis method for smart highways according to claim 1, characterized in that, The specific steps of step S5 are as follows: S51. The cloud control platform parses the global energy scheduling strategy into energy equipment control commands and sends them to the edge computing nodes corresponding to specific devices; S52. Edge computing nodes execute control commands for energy equipment, adjusting the power of charging piles and changing the working mode of energy storage systems; S53. The edge computing node collects the actual operating data of the equipment after the energy equipment control command is executed, and uses it as new local multi-source data to re-execute steps S1 to S4.

10. A comprehensive energy decision support data analysis system for smart highways, characterized in that, include: Edge computing nodes are deployed at various energy-consuming nodes along the highway to collect local multi-source data from each energy-consuming node, and perform cleaning, formatting, and preliminary local calculations to generate standardized node energy data packages. The management sub-center communicates with at least one edge computing node to aggregate the node energy data packets of each energy-consuming node and perform spatiotemporal correlation fusion with the corresponding traffic business data to extract regional energy-traffic correlation features. It then uses these regional energy-traffic correlation features to update the state of the region's digital twin model and generate a regional structured dataset. The cloud control platform communicates with the management sub-center and is used to receive regional structured datasets from at least one management sub-center. It uses a prediction model to predict energy demand and new energy power generation for future periods based on historical and real-time data. Based on the prediction results, the real-time status of the highway energy system, external boundary conditions, and carbon emission accounting rules, it performs real-time accounting and carbon flow tracking of the highway energy system's carbon emissions. It also solves an optimization decision model with the goals of economy, low carbon emissions, or high proportion of new energy consumption to generate a global energy dispatch strategy. The strategy execution and feedback module is used to distribute the global energy scheduling strategy to the corresponding management sub-center and edge computing nodes. The edge computing nodes execute the energy equipment control commands and feed back the real-time status and effect data of the highway energy system after the command execution to the cloud control platform to optimize the iterative update of the optimization decision model and prediction model.

Images