Service area new energy charging management method and system based on multi-device data analysis

Abstract

The invention relates to the technical field of charging pile power dispatching, in particular to a service area new energy charging management method and system based on multi-device data analysis. The method comprises the following steps: acquiring real-time operation monitoring parameters of all charging piles in a service area; multi-dimensional feature perception and charging behavior time popularity distribution analysis are carried out, and a dynamic charging behavior hotspot map is constructed; carrying out multi-time-point power sampling according to the dynamic charging behavior hotspot map, carrying out power cooperation path topology evolution, and constructing a power cooperation path network; performing charging pile power trend prediction and power peak congestion evolution according to the power cooperation path network, and generating a power peak congestion state feature of the charging pile; and obtaining new energy traffic flow data of the service area, carrying out unit time traffic flow average calculation, and carrying out traffic flow prediction to obtain a traffic flow prediction thermodynamic diagram. Charging power scheduling is carried out through demand prediction, and the operation efficiency and stability of charging equipment are improved.

Description

Technical Field

[0001] The present invention relates to the technical field of charging pile power scheduling, and in particular to a service area new energy charging management method and system based on multi-device data analysis. Background Art

[0002] With the rapid development and popularization of new energy vehicles, the construction of charging infrastructure has become one of the key factors to support its sustainable development. Especially in public scenarios such as highway service areas, the demand for new energy vehicle charging is characterized by high frequency and concentration. As an important node connecting transportation and energy, charging piles in service areas are gradually evolving into a core component of the smart transportation system. However, traditional charging management methods can no longer meet the current growing demand for vehicle charging and diversified service scenarios, and more efficient and intelligent management methods are urgently needed to support them.

[0003] In the current application environment, service areas are usually equipped with multiple charging devices, which are from different manufacturers and have different specifications. The data such as operating status, charging efficiency, and user behavior are isolated from each other, lacking unified data management and intelligent analysis capabilities. This "data island" phenomenon seriously restricts the operating efficiency and user experience of charging facilities in service areas. At the same time, due to the lack of real-time data analysis and multi-device coordination capabilities of traditional management methods, problems such as high equipment idle rate, serious waiting queues, uneven load distribution, and delayed power dispatch are prone to occur, affecting the overall energy utilization rate and service quality.

[0004] The current charging management system mostly relies on fixed time scheduling rules or manual maintenance plans, lacks in-depth mining and intelligent analysis of multi-device operation data, and is difficult to achieve rapid diagnosis of equipment failures, real-time optimization of charging behavior, and dynamic configuration of power resources. Therefore, there is an urgent need for a more intelligent new energy charging management method that can break through the data barriers between different devices, integrate multi-dimensional operation information, and achieve reasonable allocation of charging resources, dynamic monitoring of equipment operation status, and accurate prediction of user behavior through intelligent algorithms, thereby comprehensively improving the intelligent management level of new energy charging in service areas. Summary of the invention

[0005] In order to solve the above technical problems, the present invention proposes a service area new energy charging management method and system based on multi-device data analysis to solve at least one of the above technical problems.

[0006] To achieve the above object, the present invention provides a service area new energy charging management method based on multi-device data analysis, comprising the following steps: Step S1: Obtain real-time operation monitoring parameters of all charging piles in the service area; perform multi-dimensional feature perception and charging behavior time heat distribution analysis, and construct a dynamic charging behavior hotspot map; Step S2: Perform multi-time point power sampling according to the dynamic charging behavior hotspot map, and perform power coordination path topology evolution to build a power coordination path network; Step S3: predicting the power trend of the charging pile and the evolution of the power peak congestion according to the power cooperative path network, and generating the power peak congestion state characteristics of the charging pile; Step S4: acquiring new energy vehicle flow data of the service area, calculating the average vehicle flow per unit time, and performing vehicle flow prediction to obtain a vehicle flow prediction heat map; Step S5: performing comprehensive charging power demand prediction and pre-charging power intelligent scheduling based on the power peak congestion state characteristics according to the vehicle flow prediction heat map, and constructing a charging power intelligent scheduling strategy; Step S6: Shield the faulty charging piles in the charging power intelligent scheduling strategy, make global power rescheduling decisions, and build an intelligent charging management engine.

[0007] The present invention can establish a comprehensive perception of the operation status of charging piles in the service area by collecting the operation status data of all charging piles in real time (such as power, current, usage frequency, queuing status, etc.). Through heat distribution, time period analysis and other means, the user's high-frequency charging period and spatial concentration area can be identified. Provide accurate spatiotemporal data basis for subsequent load prediction and resource allocation, and improve overall response efficiency. Through multi-time point power sampling, identify the collaborative relationship and path evolution pattern between different charging piles in power load. Construct the dispersed devices in the physical space into a collaborative network through the logical relationship on the data, and provide graph structure support for scheduling optimization. The network topology can be used for subsequent load peak prediction, congestion trend simulation, etc., to provide structural input for trend analysis. It can identify in advance which charging piles may have peak load and queue congestion in the future time period. The extracted "peak congestion state characteristics" can be used as model input to improve the pertinence of subsequent scheduling. Achieve early perception of future load pressure, which is convenient for the scheduling system to take early peak regulation, guidance and other control measures. The traffic flow prediction establishes a dynamic connection between traffic flow and charging demand to avoid the situation of "some people without piles" or "there are piles but no people". The heat map visually displays the traffic concentration areas and time periods in the service area, which is conducive to the reallocation of resource space. Before formulating a dispatch plan, understand the traffic flow direction and peak hours to make power dispatch more forward-looking and accurate. Integrate the traffic flow prediction heat map with the congestion state characteristics to achieve a more accurate prediction of charging power demand. Rationally allocate power resources in the dispatch strategy, guide vehicles to use idle charging piles, and alleviate pressure in high-load areas. Effectively reduce charging waiting time, improve user charging experience, and ensure equipment operation efficiency and grid stability. By shielding the faulty pile nodes, avoid them from interfering with the dispatch strategy of the entire system. If some piles are unavailable, the system can reconstruct the dispatch plan in real time based on the remaining resources and traffic forecast results. Form an intelligent charging management center that integrates data perception, prediction analysis, strategy generation, dispatch execution, and fault tolerance.

[0008] In this specification, a service area new energy charging management system based on multi-device data analysis is provided, which is used to execute the service area new energy charging management method based on multi-device data analysis as described above, including: The multi-dimensional feature perception module is used to obtain the real-time operation monitoring parameters of all charging piles in the service area; and to perform multi-dimensional feature perception and charging behavior time heat distribution analysis to build a dynamic charging behavior hotspot map; The power collaborative path module is used to perform multi-time point power sampling according to the dynamic charging behavior hot spot map, and to perform power collaborative path topology evolution to build a power collaborative path network; The power trend prediction module is used to predict the power trend of charging piles and the evolution of power peak congestion according to the power cooperative path network, and generate the power peak congestion state characteristics of the charging piles; The vehicle flow prediction module is used to obtain the new energy vehicle flow data in the service area, calculate the average vehicle flow per unit time, and predict the vehicle flow to obtain a vehicle flow prediction heat map; A power intelligent dispatching module, used to perform comprehensive charging power demand prediction and pre-charging power intelligent dispatching based on the power peak congestion state characteristics according to the traffic flow prediction heat map, and build a charging power intelligent dispatching strategy; The fault shielding module is used to shield faulty charging piles for the charging power intelligent scheduling strategy, make global power rescheduling decisions, and build an intelligent charging management engine.

[0009] The present invention can timely reflect the working status and usage of charging piles by acquiring real-time operation monitoring parameters of all charging piles, which provides an accurate data basis for subsequent analysis and decision-making. Multi-dimensional feature perception and time heat distribution analysis of charging behavior can identify the peak time and hot spots of charging behavior, which helps to understand the charging habits and needs of users and optimize the configuration of charging resources. A dynamic charging behavior hot spot map is constructed so that managers can intuitively understand the usage of charging piles and support decision-making and resource allocation. Multi-point power sampling is performed according to the dynamic charging behavior hot spot map, which can capture the power changes in different time periods and provide a basis for power management. The power collaborative path topology evolution is performed to construct a power collaborative path network, which can help identify the power relationship between charging piles and support more efficient power scheduling and management. The power trend prediction of charging piles through the power collaborative path network can identify the changes in power demand in advance and help managers effectively deal with potential power peak congestion problems. The power peak congestion state characteristics of the charging piles are generated, which enables managers to quickly identify and take measures to avoid the decline in user experience due to insufficient power. Obtaining the traffic data of new energy vehicles in the service area and calculating the average traffic volume per unit time can provide support for charging demand prediction. Through traffic flow prediction, a traffic flow prediction heat map is generated, so that managers can intuitively understand the changes in traffic flow in different time periods, thereby optimizing the distribution and scheduling of charging piles. According to the traffic flow prediction heat map, a comprehensive charging power demand prediction is made for the power peak congestion state characteristics, which helps to reasonably allocate charging resources and reduce peak loads. Constructing a charging power intelligent scheduling strategy can reasonably schedule charging power in different time periods, improve the utilization rate of charging piles, and reduce user waiting time. The charging power intelligent scheduling strategy shields faulty charging piles to ensure that when a charging pile fails, the system can automatically adjust the power allocation to minimize user inconvenience. Make global power rescheduling decisions and build an intelligent charging management engine so that the entire system can quickly adapt to dynamic environments and ensure the stability and continuity of charging services. BRIEF DESCRIPTION OF THE DRAWINGS

[0010] Figure 1 A schematic diagram of the steps of a service area new energy charging management method based on multi-device data analysis according to the present invention; Figure 2 Detailed implementation flow chart of step S1; Figure 3 Detailed implementation flow chart of step S2; Figure 4 Detailed implementation flow chart of step S3. DETAILED DESCRIPTION

[0011] It should be understood that the specific embodiments described herein are only used to explain the present invention, and are not used to limit the present invention.

[0012] The example of this application provides a service area new energy charging management method and system based on multi-device data analysis. The execution subject of the service area new energy charging management method and system based on multi-device data analysis includes but is not limited to: mechanical equipment, data processing platform, cloud server node, network upload device, etc. equipped with the system can be regarded as the general computing node of this application, and the data processing platform includes but is not limited to: at least one of an audio image management system, an information management system, and a cloud data management system.

[0013] See also Figures 1 to 4 The present invention provides a service area new energy charging management method based on multi-device data analysis, and the service area new energy charging management method based on multi-device data analysis includes the following steps: Step S1: Obtain real-time operation monitoring parameters of all charging piles in the service area; perform multi-dimensional feature perception and charging behavior time heat distribution analysis, and construct a dynamic charging behavior hotspot map; Step S2: Perform multi-time point power sampling according to the dynamic charging behavior hotspot map, and perform power coordination path topology evolution to build a power coordination path network; Step S3: predicting the power trend of the charging pile and the evolution of the power peak congestion according to the power cooperative path network, and generating the power peak congestion state characteristics of the charging pile; Step S4: acquiring new energy vehicle flow data of the service area, calculating the average vehicle flow per unit time, and performing vehicle flow prediction to obtain a vehicle flow prediction heat map; Step S5: performing comprehensive charging power demand prediction and pre-charging power intelligent scheduling based on the power peak congestion state characteristics according to the vehicle flow prediction heat map, and constructing a charging power intelligent scheduling strategy; Step S6: Shield the faulty charging piles in the charging power intelligent scheduling strategy, make global power rescheduling decisions, and build an intelligent charging management engine.

[0014] The present invention can establish a comprehensive perception of the operation status of charging piles in the service area by collecting the operation status data of all charging piles in real time (such as power, current, usage frequency, queuing status, etc.). Through heat distribution, time period analysis and other means, the user's high-frequency charging period and spatial concentration area can be identified. Provide accurate spatiotemporal data basis for subsequent load prediction and resource allocation, and improve overall response efficiency. Through multi-time point power sampling, identify the collaborative relationship and path evolution pattern between different charging piles in power load. Construct the dispersed devices in the physical space into a collaborative network through the logical relationship on the data, and provide graph structure support for scheduling optimization. The network topology can be used for subsequent load peak prediction, congestion trend simulation, etc., to provide structural input for trend analysis. It can identify in advance which charging piles may have peak load and queue congestion in the future time period. The extracted "peak congestion state characteristics" can be used as model input to improve the pertinence of subsequent scheduling. Achieve early perception of future load pressure, which is convenient for the scheduling system to take early peak regulation, guidance and other control measures. The traffic flow prediction establishes a dynamic connection between traffic flow and charging demand to avoid the situation of "some people without piles" or "there are piles but no people". The heat map visually displays the traffic concentration areas and time periods in the service area, which is conducive to the reallocation of resource space. Before formulating a dispatch plan, understand the traffic flow direction and peak hours to make power dispatch more forward-looking and accurate. Integrate the traffic flow prediction heat map with the congestion state characteristics to achieve a more accurate prediction of charging power demand. Rationally allocate power resources in the dispatch strategy, guide vehicles to use idle charging piles, and alleviate pressure in high-load areas. Effectively reduce charging waiting time, improve user charging experience, and ensure equipment operation efficiency and grid stability. By shielding the faulty pile nodes, avoid them from interfering with the dispatch strategy of the entire system. If some piles are unavailable, the system can reconstruct the dispatch plan in real time based on the remaining resources and traffic forecast results. Form an intelligent charging management center that integrates data perception, prediction analysis, strategy generation, dispatch execution, and fault tolerance.

[0015] In the embodiment of the present invention, refer to Figure 1 , is a schematic flow chart of the steps of a service area new energy charging management method based on multi-device data analysis of the present invention. In this example, the steps of the service area new energy charging management method based on multi-device data analysis include: Step S1: Obtain real-time operation monitoring parameters of all charging piles in the service area; perform multi-dimensional feature perception and charging behavior time heat distribution analysis, and construct a dynamic charging behavior hotspot map; In this embodiment, the real-time monitoring system of all charging piles in the service area is determined to ensure that the necessary operating parameters can be extracted from them. These parameters usually include charging pile ID, current voltage, current, charging status (idle, charging, fault), number of users, charging time, etc. Set to obtain data from the monitoring system every 5 minutes to ensure the timeliness and accuracy of the data. Through the API interface or data acquisition tool, extract real-time data from the monitoring system of all charging piles at regular intervals and store it in the database. Ensure that the data format is consistent for subsequent analysis. Create a database table containing fields such as "charging pile ID", "current voltage", "current current", "charging status", "number of users" and "timestamp" to facilitate subsequent data query and analysis. Verify the acquired real-time data to ensure that there are no duplicates, missing or abnormal values. Remove records that are beyond the reasonable range (such as voltage outside the normal range) to maintain data integrity. Set the voltage range to 100V to 400V, and any value outside this range should be marked as abnormal and recorded for subsequent analysis. Based on the collected real-time data, extract multi-dimensional features, including but not limited to voltage fluctuations, charging power, changes in the number of users, charging time, etc. of the charging piles. These features will be used for subsequent dynamic analysis. Calculate the average voltage and current of each charging pile in the past hour, and record the changes in the number of users in each time period. Aggregate the extracted feature data and summarize them by time period (such as every 30 minutes) to form a comprehensive feature matrix. Ensure that the feature data in each time period can reflect the dynamic state of the charging pile. For each charging pile, count its average voltage, average current and number of charging users in every 30 minutes for subsequent analysis. Apply data analysis techniques (such as time series analysis or cluster analysis) to analyze the multi-dimensional features and identify the dynamic change pattern of charging behavior. This process helps to understand the usage of charging piles and user behavior. Use the time series analysis method to identify the peak charging period and its corresponding charging pile usage, and record the analysis results. Based on the aggregated feature data, calculate the popularity of charging behavior in different time periods, which can be comprehensively scored by indicators such as the number of users, charging time and charging power to identify the time periods with high popularity. Set the heat score formula as: heat = number of users × average charging power × charging time, and calculate the heat value for each time period. Visualize the calculated heat value and build a dynamic charging behavior heat map. The heat map should be able to intuitively display the usage of charging piles in different time periods to help managers quickly identify peak hours. Use the heat map tool to map the heat value to the timeline, and use different colors to indicate the heat level, which is convenient for managers to make decisions and allocate resources.

[0016] Step S2: Perform multi-time point power sampling according to the dynamic charging behavior hotspot map, and perform power coordination path topology evolution to build a power coordination path network; In this embodiment, based on the dynamic charging behavior heat map constructed in the previous step, the peak usage period of the charging pile is identified, which usually corresponds to the time period when the user's charging demand is most concentrated. Ensure that the heat map can accurately reflect the dynamic changes of charging behavior. Assuming that the heat map shows that 08:00-09:00 and 18:00-19:00 every day are peak periods, these periods are used as the key periods for power sampling. During the identified peak period, power data is collected from each charging pile real-time monitoring system. These data should include information such as the actual charging power, charging status and number of users of each charging pile at different time points. Set the power data to be collected once a minute, and record the power value of each charging pile during the peak period. Assume that during 08:00-09:00, the power of charging pile A is recorded as 50 kW, the power of charging pile B is 60kW, etc. Graph theory methods are used to analyze the power coordination path between charging piles. Each charging pile is regarded as a node in the graph, and the power flow relationship is regarded as the edge between the nodes. Select a suitable algorithm (such as Dijkstra algorithm or minimum spanning tree algorithm) to build the network. Set the power flow relationship between adjacent charging piles as the weight of the edge. If the power flow of charging pile A during peak hours affects charging pile B, a connection is established in the graph. Generate a power collaborative path network for charging piles based on the collected power data. Identify the relationships between charging piles, determine which charging piles affect each other during peak hours, and the direction of their power flow. If the frequency of use of charging pile A affects the power demand of charging pile B during peak hours, establish a connection in the network and record its power flow information. After building the basic power collaborative path network, perform topological evolution and dynamically adjust the network structure based on changes in real-time power data. Ensure that the network can adapt to changes in charging demand in different time periods. If the power demand of charging pile C increases sharply during a certain time period, adjust the connection relationship in the network to ensure that charging pile C can obtain sufficient power support and record this change. Record the constructed power collaborative path network and its topological evolution results in the database, and use visualization tools to display the network structure. Ensure that managers can intuitively see the collaborative relationship and power flow between charging piles. Generate a network diagram showing the connections between chargers and the direction of power flow to help managers identify charging capacity and potential bottlenecks during peak hours.

[0017] Step S3: predicting the power trend of the charging pile and the evolution of the power peak congestion according to the power cooperative path network, and generating the power peak congestion state characteristics of the charging pile; In this embodiment, based on the previous power collaborative path network, the power data of the current charging pile is collected. These data should include historical charging power, number of users, charging time and other characteristics, so as to predict the power trend.

[0018] Set the data collection cycle to the hourly power data for the past week and extract relevant features. Ensure that the records of each charging pile include timestamp, average power, maximum power, and number of charging users.

[0019] Select a forecasting model: Select a suitable power trend prediction model based on the data characteristics. Common models include time series prediction models (such as ARIMA, SARIMA) and machine learning models (such as LSTM, random forest). The selected model should be able to handle the characteristics of time series data.

[0020] The LSTM model was chosen because it can capture long-term dependencies when processing time series data and is suitable for predicting the power change trend of charging piles.

[0021] Model training and validation: Use historical data to train the selected prediction model. Divide the data into training and test sets to ensure that the model can learn the changing pattern of charging pile power and verify it on the test set.

[0022] The charging pile power data from the past two weeks is used for training to ensure that the root mean square error (RMSE) of the model during prediction is lower than the set threshold to ensure the accuracy of the prediction.

[0023] To perform power trend forecasting: The trained model is used to predict the charging pile power in future time periods, and the predicted value for each time period is recorded, which will provide basic data for the subsequent evolution of power peak congestion.

[0024] Predict the power demand of each charging pile in the next 24 hours and record the predicted power value for each time period. Based on the power trend prediction results, identify the power peak congestion that may occur in the charging pile during peak hours. Set the congestion threshold. When the demand of a charging pile exceeds its maximum power carrying capacity, it is considered congested. Set the maximum power of the charging pile to 100 kW. If it is predicted that the demand of a charging pile reaches 120 kW in a certain time period, it is marked as a congestion state. Use the time series analysis method to dynamically monitor the changes in the power demand of the charging pile and analyze the changes in the congestion state in different time periods. Record the congestion of each charging pile and its evolution trend. If the power demand of charging pile A increases from 80 kW to 120 kW during the peak period (such as 18:00-19:00), record this evolution process and analyze the possible reasons (such as concentrated charging by users). Summarize the identified congestion states and generate the power peak congestion state features of the charging pile. These features include information such as the time when the congestion occurred, the duration, the charging pile ID involved and its power demand. Record which charging piles are congested during a specific peak period, and summarize their common characteristics, such as the average congestion duration of 30 minutes and the number of charging piles involved.

[0025] Step S4: acquiring new energy vehicle flow data of the service area, calculating the average vehicle flow per unit time, and performing vehicle flow prediction to obtain a vehicle flow prediction heat map; In this embodiment, the source of the new energy vehicle flow data is determined, which usually includes sensors, cameras or traffic monitoring systems installed in the service area. These devices should be able to monitor the number of new energy vehicles entering and leaving the service area in real time. Set to obtain traffic flow data from the traffic monitoring system every 5 minutes, and ensure that the equipment works normally to ensure the accuracy of the data. Through the API interface or data acquisition tool, the new energy vehicle flow data is extracted from the monitoring system regularly and stored in the database. Ensure that the data format is consistent for subsequent processing. Create a database table containing fields such as "timestamp", "vehicle type", "entry and exit direction", "traffic flow", etc., for subsequent analysis and query. During the data collection process, the acquired traffic flow data is verified. Check whether there are missing values, duplicate records or outliers, and perform necessary cleaning processing. Set a reasonable traffic flow range to filter out unreasonable data (such as negative values ​​or the number of vehicles exceeding the maximum carrying capacity). If the traffic flow in a certain time period exceeds expectations, it is marked as abnormal and recorded. According to the analysis requirements, the time is divided into multiple windows (such as every hour or every 30 minutes) to facilitate segmented statistics of traffic flow data. Ensure that the data in each time window is complete. Set every 30 minutes as a time window and record the traffic flow data in each time period. Calculate the average of the traffic flow data per unit time in each time window. The traffic flow per unit time is obtained by dividing the traffic flow in each window by the window length. For a 30-minute time window, if the total traffic flow recorded in the window is 300 vehicles, the traffic flow per unit time is 300 vehicles / 0.5 hours = 600 vehicles / hour. According to the characteristics of the traffic flow data per unit time, select a suitable traffic flow prediction model. Commonly used models include ARIMA model, time series analysis, and machine learning models (such as LSTM). The selected model should be able to handle the characteristics of time series data. Select the ARIMA model because it performs well in handling seasonal and trend data and is suitable for traffic flow prediction. Use historical traffic flow data to train the selected prediction model. Divide the data into training set and test set to ensure that the model can learn the change pattern of traffic flow, and verify it on the test set to evaluate the accuracy of the prediction. Use the hourly traffic data from the past three months for training, and ensure that the root mean square error (RMSE) of the model during prediction is lower than the set threshold. Use the trained model to predict the traffic volume in future time periods, generate traffic volume prediction values, and record the prediction results for each time period. Predict the changes in traffic volume in the next 24 hours, and record the prediction values ​​for each time period. For example, it is expected that the traffic volume in a certain time period will be 500 vehicles. Arrange the predicted traffic volume data into a format suitable for heat map display. This can be done by combining the time period and the corresponding predicted traffic volume to construct a matrix. Create a two-dimensional array with rows representing time periods and columns representing traffic volume for subsequent visualization.Use data visualization tools (such as Tableau and Matplotlib) to generate heat maps from the sorted traffic flow forecast data. Ensure that the heat map can intuitively display the traffic flow distribution in different time periods. Different shades of color are used in the heat map to indicate the high and low traffic flow, helping managers quickly identify peak and low periods.

[0026] Step S5: performing comprehensive charging power demand prediction and pre-charging power intelligent scheduling based on the power peak congestion state characteristics according to the vehicle flow prediction heat map, and constructing a charging power intelligent scheduling strategy; In this embodiment, the traffic flow prediction heat map and the previously extracted power peak congestion state characteristics are combined to integrate data for comprehensive charging power demand prediction. It is necessary to ensure that the data covers the actual power demand of the charging pile and the future traffic flow changes. Set the data time period to the next 24 hours, and record the predicted traffic flow, historical charging power data and the current available status of the charging pile in each time window. According to the characteristics of the integrated data, select a suitable prediction model for comprehensive charging power demand prediction. Common models include linear regression, decision tree, random forest, etc. The selected model should be able to handle multi-dimensional feature data and consider the relationship between traffic flow and charging power. The random forest model is selected because it performs well in handling complex feature relationships and can effectively handle nonlinear relationships. Use historical data to train the selected prediction model. Divide the data into a training set and a test set to ensure that the model can learn the changing pattern of charging demand, and verify it on the test set to evaluate the accuracy of the prediction. Use the traffic flow and charging power data of the past three months for training to ensure that the root mean square error (RMSE) of the model in prediction is lower than the set threshold (such as 15%). Use the trained model to predict the comprehensive charging power demand in the future time period, and record the predicted power demand value for each time period, which will provide basic data for subsequent intelligent scheduling decisions. Predict the charging power demand in the next 24 hours, and it is expected that the charging demand will reach 100 kW during certain peak hours (such as 08:00-09:00). Design a charging power intelligent scheduling strategy based on the predicted comprehensive charging power demand to ensure that charging resources can be effectively allocated during high-demand periods. The scheduling strategy should consider the maximum load capacity of the charging pile, the user queue situation, and the priority of charging demand. Set the maximum power of the charging pile to 100 kW during the peak period. If the charging demand is predicted to exceed this value, it is necessary to give priority to the power allocation to the user with the longest queue time. Use a dynamic scheduling algorithm (such as the optimal allocation algorithm or the greedy algorithm) to schedule the charging pile in real time, and dynamically adjust the charging power allocation according to the predicted demand. Ensure that the charging pile is used most efficiently during the peak charging demand period. If the charging demand is predicted to be 150 kW during 08:00-09:00, and there are two charging piles, the power can be allocated to each charging pile as 75 kW to avoid overloading a single charging pile. After implementing the scheduling strategy, monitor the scheduling effect in real time, collect user feedback and charging pile operation data. Ensure that the scheduling strategy can effectively improve the charging efficiency, and adjust the strategy in time to cope with changes. Monitor the user's waiting time and charging success rate. If it is found that the user's waiting time is too long in a certain period of time, it is necessary to re-evaluate the charging power allocation strategy.

[0027] Step S6: Shield the faulty charging piles in the charging power intelligent scheduling strategy, make global power rescheduling decisions, and build an intelligent charging management engine.

[0028] In this embodiment, the real-time monitoring system of the integrated charging pile automatically identifies the fault status of the charging pile by monitoring parameters (such as voltage, current, temperature, etc.) and fault record data. Establish fault identification rules for real-time detection. Set fault standards: If the current fluctuation of a charging pile exceeds ±20% or the temperature exceeds 80°C within a certain period of time, it is marked as a faulty pile. Once a faulty charging pile is identified, its status is immediately updated to "fault", and the time of the fault, the charging pile ID and the fault type are recorded. These data will be used for the adjustment of subsequent scheduling strategies. If the charging pile ID is "C001" and is judged to be faulty at 08:15, it is marked, and the cause of the fault (such as overheating) and its impact range are recorded. A shielding mechanism for faulty charging piles is implemented in the intelligent scheduling strategy for charging power to ensure that these charging piles are not considered in the scheduling decision, so as to avoid allocating faulty charging piles to users and reduce the charging risk of users. Update the scheduling algorithm so that it excludes faulty piles when calculating available charging piles, thereby reallocating power to other normally operating charging piles. Analyze the global power scheduling strategy based on the current usage status and fault conditions of the charging pile. Ensure that charging resources can be efficiently allocated after the faulty charging pile is shielded. If there are 4 charging piles available but 1 fails, the power allocation capacity of the remaining 3 charging piles needs to be re-evaluated. Use dynamic scheduling algorithms (such as optimal allocation algorithm, greedy algorithm or linear programming) for global power rescheduling. Dynamically adjust the power allocation plan according to the real-time charging demand and available charging piles. If the predicted total charging demand is 150 kW during peak hours and the total power of available charging piles is 200 kW, power can be allocated preferentially according to the user's queue situation. While rescheduling is being executed, establish a real-time monitoring and feedback mechanism to ensure that the scheduling effect and user feedback can be obtained in a timely manner. Dynamically adjust the scheduling strategy according to the user's charging status and waiting time. Monitor the actual waiting time of users and the usage rate of charging piles in a certain period of time. If it is found that the usage rate of a charging pile is too high and the user's waiting time is prolonged, its power allocation needs to be readjusted. Record the results of the global scheduling decision in the database and generate a detailed scheduling report. The report should include the execution of the scheduling strategy, user feedback and the efficiency of the charging piles to facilitate subsequent analysis and optimization. Generate a report containing "time period", "charging demand forecast", "actual power distribution", and "user waiting time" to help managers evaluate the effectiveness of the scheduling strategy and make necessary adjustments. Design the architecture of the intelligent charging management engine to ensure that it can process monitoring data, fault identification, scheduling decisions, and user feedback in real time. The engine should include a data input module, a fault detection module, a scheduling decision module, and a data output module. Ensure that the engine can accept data streams from charging piles and can output scheduling results and user notifications in real time. Integrate the various modules to ensure smooth data flow, and perform functional tests with actual data to verify the effectiveness and stability of the engine.Use simulated data for testing to observe the engine's ability to respond to different failure situations and ensure that it can quickly adjust its scheduling strategy.

[0029] In this embodiment, refer to Figure 2 , is a flowchart of detailed implementation steps of step S1. In this embodiment, the detailed implementation steps of step S1 include: Obtaining real-time operation monitoring parameters of all charging piles in the service area; calculating the voltage and current parameters of the charging piles according to the real-time operation monitoring parameters of the charging piles to obtain the electrical characteristics of each charging pile; The usage frequency of the real-time operation monitoring parameters of the charging pile is successively counted to extract the usage frequency parameters of the charging pile; Extracting the charging time and idle time period of the real-time operation monitoring parameters of the charging pile; Perform multi-dimensional feature perception on the usage frequency parameters of the electrical characteristic charging pile, the charging time and the idle period, and construct a multi-dimensional feature matrix; Perform time stamp synchronization on the multi-dimensional feature matrix, perform time vectorization decomposition, and construct a rasterized time window for charging behavior; Each charging behavior is extracted according to the rasterized time window of charging behavior, and the time heat distribution analysis of charging behavior is performed to construct a dynamic charging behavior hotspot map.

[0030] In this embodiment, the source of real-time monitoring data of the charging pile is determined, which usually includes parameters such as voltage, current, charging status, and frequency of use. These data can be obtained through the communication interface of the charging pile (such as REST API, MQTT protocol, etc.). Set the real-time data of the charging pile to be collected every 30 seconds to ensure the timeliness and accuracy of the data. By configuring the interface, obtain the unique identifier of each charging pile for subsequent data processing and analysis. By writing a data acquisition program, the real-time operation monitoring parameters of the charging pile are obtained regularly and the data is stored in the database. Ensure that the data format is consistent for subsequent analysis. Create a database table with storage fields including "charging pile ID", "voltage", "current", "charging status", "timestamp", etc. to ensure that this information can be recorded for each data collection. Verify the acquired real-time monitoring parameters to ensure the integrity and validity of the data. Remove unreasonable abnormal values ​​(such as voltage less than 0 or current exceeding the rated range). Set the normal voltage range to 100V to 400V and the normal current range to 0A to 100A. Any value outside this range should be marked as abnormal and recorded. Calculate the electrical characteristics of each charging pile based on the collected real-time monitoring parameters. Electrical characteristics include charging power (P=U*I), energy consumption, etc. Set the voltage of the charging pile to 220V and the current to 16A, then the charging power is 220V * 16A =3520W, record the power value and its timestamp information. Calculate the energy consumption of each charging pile based on the charging time and charging power. Use the formula: Energy (kWh) = Power (kW) * Time (h). When charging lasts for 1 hour, the energy consumption is 3520W / 1000 * 1 = 3.52 kWh, store this information for subsequent analysis. According to the real-time monitoring parameters of the charging pile, count the usage frequency of each charging pile, that is, the number of times the charging pile is used within a certain period of time. Set the statistical period (such as daily or hourly). Calculate the number of times each charging pile has been charged in the past 24 hours and record the usage frequency of each charging pile. Summarize the statistical results into the database to form a report on the usage frequency of the charging piles. And sort them according to the usage frequency to identify the most popular charging piles. Generate a table containing the charging pile ID and the corresponding usage frequency to record the usage of the charging pile for subsequent analysis. Through the time series analysis method, analyze the changing trend of the usage frequency of the charging piles and identify the peak usage periods and trough periods. Use the moving average method to calculate the usage frequency per hour and form a trend chart for visual analysis. According to the real-time monitoring parameters, extract the start time and end time of each charging and calculate the charging duration. Record the duration information of each charging session. If a charging pile starts charging at 08:00 and ends at 09:00, the charging duration is 1 hour, and store the duration and related time information. Based on the charging duration, identify the idle period of the charging pile.The idle period is defined as the time period that is not used, and the idle period of each charging pile is recorded. If a charging pile is not used between 09:00 and 10:00, this period is recorded as the idle period. The extracted electrical characteristics, usage frequency, charging time and idle period of the charging pile are integrated into a multidimensional feature matrix. Each row represents a charging pile, and each column represents a feature. The columns of the feature matrix can include "charging pile ID", "voltage", "current", "usage frequency", "charging time", "idle period", etc. to ensure the comprehensiveness of the data. The feature matrix is ​​standardized to eliminate the dimensional influence between different features and ensure the accuracy of subsequent analysis. The Z-score standardization method is used to adjust the mean of each feature to 0 and the standard deviation to 1 to ensure data adaptability. The timestamps in the multidimensional feature matrix are synchronized to ensure that all feature data are analyzed under the same time base and eliminate the influence of time differences. The timestamps of all charging piles are unified to UTC time, and the data of different time periods are interpolated to ensure time continuity. Vectorize the timestamp data to form time series data for subsequent analysis. Time can be divided into fixed time windows (such as every 5 minutes or every hour). Divide a day into 288 5-minute time windows, and summarize the charging behavior data in each time window. Record the processed time window data in the database to form a time series of charging behavior for subsequent analysis and visualization. Generate a time window data table to record the number of charging times and total charging time in each time window for subsequent heat distribution analysis. According to the time window data, extract each charging behavior, including charging start time, end time, charging pile ID and other information for subsequent analysis. Record the charging pile ID and charging time in each time window to form a charging behavior record table. Perform heat distribution analysis on the extracted charging behavior data, and calculate the heat value based on the frequency of use and charging time of the charging pile. The heat value can be a weighted value of the number of charging times. Set the heat value calculation formula as: heat value = frequency of use × average charging time, and record the heat value of each charging pile. According to the calculated heat value, construct a dynamic charging behavior hotspot map. Use visualization tools to display the heat distribution of different charging piles to facilitate managers' decision-making. Generate a heat map to display the heat distribution of each charging pile, using different colors to indicate the heat level, to help identify the efficiency of charging piles and room for improvement.

[0031] In this embodiment, refer to Figure 3 , is a flowchart of detailed implementation steps of step S2. In this embodiment, the detailed implementation steps of step S2 include: Calculate the power of each charging behavior based on the dynamic charging behavior hotspot map to generate the power characteristics of each charging pile; Performing power sampling at multiple time points on the charging power characteristics to construct charging power curves for multiple charging piles; Perform discrete Fourier transform on the charging power curves of multiple charging piles to generate charging power frequency domain data; Perform spectrum peak detection on the charging power frequency domain data and mark multiple spectrum peaks; Identify the peak timing shift characteristics of the spectrum peak and mark the high-frequency charge and discharge fluctuation nodes; The power collaborative path topology evolves according to the high-frequency charging and discharging fluctuation nodes, and constructs a power collaborative path network.

[0032] In this embodiment, the charging behavior record of each charging pile is extracted from the previously generated dynamic charging behavior hotspot map, including the charging start time, end time, charging pile ID and other information, which will be used to calculate the power of each charge. The charging start time of a charging pile is recorded as 08:00, the end time is 09:00, the charging power is 3.5 kW, and the energy consumption of the charging session is calculated to be 3.5 kW × 1 h = 3.5 kWh. According to the extracted charging behavior data, the charging power characteristics of each charging pile are calculated one by one. The power characteristic calculation should take into account the voltage and current parameters of different charging periods, and the formula P = U × I is usually used. If the voltage of a charging pile at 08:30 is 220V and the current is 16A, the charging power at this moment is calculated to be 220V × 16A = 3520W. The power characteristics of each charging session and its timestamp are recorded. The calculated charging power characteristics of each time are stored in the database to ensure data structure and facilitate subsequent analysis. Each record should contain information such as the charging pile ID, charging time, and power characteristics. Create a database table with storage fields including "charging pile ID", "charging time", "charging power", etc. for subsequent query and analysis. Extract the power data of multiple charging piles at different time points from the stored charging power features to form a power data set. Ensure that the data covers different time periods and charging piles to obtain comprehensive power change information. Set the time period to one day, and extract the power data of each charging pile every hour, and record the power characteristics of different charging piles at different times of the day. Sort the extracted power data according to chronological order, and generate the charging power curve of each charging pile. These curves should be able to show the trend of power change over time. Use a line chart to show the power change of each charging pile within 24 hours, with time on the X axis and power on the Y axis, to ensure that the charging peak and trough can be clearly observed. Perform discrete Fourier transform (DFT) on the charging power curve data of multiple charging piles to analyze the frequency domain characteristics of charging power. DFT can convert time domain signals into frequency domain signals to reveal hidden frequency components. Set the length of the power curve data of each charging pile to 1024 points to ensure that the frequency domain characteristics can be effectively captured after the transformation. Use DFT to transform the power curve of each charging pile to generate frequency domain data, which will show the performance of each frequency component during the charging process. Use the standard DFT algorithm to transform the power data of each charging pile, and record the amplitude and phase information of each frequency point for subsequent analysis. Select an appropriate spectrum peak detection method, usually using a threshold method or a derivative-based peak detection algorithm to identify significant peaks in the spectrum, which represent the key frequency components in the charging power change. Set the peak detection threshold to an amplitude greater than 0.5 to ensure that only significant spectrum peaks are marked. Perform peak detection on the frequency domain data of each charging pile and mark all significant peaks in the spectrum and their corresponding frequencies.Record the amplitude and position of each peak for subsequent analysis. If a peak with a frequency of 60 Hz and an amplitude of 10 dB is detected in the spectrum of a charging pile, record this information. Perform time-series offset feature analysis on the detected spectrum peaks to identify the changing trend of the spectrum peaks during the charging process. This process will reveal the dynamic characteristics of the charging behavior. Observe the changes in the peaks in the time series, record the time when each peak occurs and its changes, and identify potential charging and discharging fluctuation nodes. Mark high-frequency charging and discharging fluctuation nodes, which usually indicate power fluctuations of the charging pile at high frequencies, which may be related to changes in charging demand or the working status of the equipment. If the peaks at 60 Hz and its multiples are found to change frequently in the spectrum, mark these nodes as high-frequency charging and discharging fluctuation nodes, and record their occurrence time and amplitude. Based on the identified high-frequency charging and discharging fluctuation nodes, analyze the power coordination paths between multiple charging piles. Identify the mutual influence and collaborative working relationship of different charging piles during the charging process. If the usage frequency of a charging pile is similar to that of other charging piles during peak hours, it is considered that these charging piles have a collaborative relationship, and their collaborative path characteristics are recorded. Integrate the identified collaborative path information to construct a power collaborative path network between charging piles. The network should reflect the relationship and interaction between charging piles. Use graph theory to construct a network diagram, with nodes representing charging piles and edges representing power collaborative relationships, to ensure that the network can intuitively reflect the collaborative characteristics of each charging pile. Analyze the constructed power collaborative path network, identify key nodes and edges in the network, and make optimization suggestions to improve the overall performance of charging piles. Analyze the connectivity of the network, identify the most active charging piles, and provide suggestions to optimize the layout and configuration of charging piles to improve charging efficiency.

[0033] In this embodiment, the specific steps of performing power cooperative path topology evolution according to high-frequency charging and discharging fluctuation nodes and constructing a power cooperative path network are as follows: Multi-cycle trend encoding is performed on high-frequency charging and discharging fluctuation nodes to generate charging power fluctuation change chains for different charging piles; Perform power fluctuation coupling detection on the charging power fluctuation chain, and extract and identify the power fluctuation coupling relationship between charging piles; The power fluctuation coupling relationship is mined to analyze the power correlation changes of multiple charging piles to generate the power linkage law between charging piles; Identify nearby associated charging piles based on the charging power fluctuation chain; According to the power linkage law between charging piles, the potential power coordination path of the adjacent associated charging piles is analyzed to extract the power coordination path of each charging pile; Analyze the power conflict situation of the power linkage between charging piles to obtain the power conflict situation between charging piles; Predict power flow based on power conflict between charging piles; Based on the power flow direction, the power cooperative path of each charging pile is topologically evolved to construct a power cooperative path network.

[0034] In this embodiment, relevant data is extracted from the previously identified high-frequency charging and discharging fluctuation nodes, including the time, amplitude and frequency characteristics of each node, which will be used to analyze the power fluctuation changes of the charging pile. Record the charging and discharging fluctuations of multiple charging piles in different time periods to ensure that the data covers a certain time period (such as 24 hours) and contains the specific timestamp of each node. The high-frequency fluctuation nodes are encoded using a trend analysis method, usually using a moving average method or an exponential smoothing method to capture the trend and periodic changes of the fluctuations. Set the window size to 6 hours, calculate the power fluctuation trend of each charging pile in the time window by the moving average method, and generate the encoded fluctuation data. Based on the encoded fluctuation data, generate a charging power fluctuation change chain for each charging pile. The chain should be able to reflect the changing trend and fluctuation characteristics of the charging power over time. Record the trend value of each charging pile in a specific time period and form a time series to ensure that the data can clearly show the dynamic characteristics of the power change. Select a suitable coupling detection method, such as mutual information or correlation coefficient analysis, to identify the power fluctuation coupling relationship between charging piles, which can quantify the mutual influence between charging piles. Set the threshold of mutual information to 0.5 to ensure that only obvious fluctuation coupling relationships are detected. The coupling detection is performed on the fluctuation change chain of different charging piles, the coupling relationship between each pair of charging piles is calculated, and the charging piles with significant coupling relationship are identified. The coupling coefficient between charging pile A and charging pile B is recorded. If its value exceeds the set threshold, it is considered that the two charging piles have a power fluctuation coupling relationship. Cluster analysis or association rule mining methods are used to analyze the power fluctuation coupling relationship between charging piles and extract the power linkage law between charging piles. The K-means clustering algorithm is used to cluster the power coupling data to identify the linkage characteristics of charging piles. By analyzing the clustering results, the power correlation changes between charging piles are identified, and the power linkage law of charging piles in different time periods is extracted. If it is found that the usage frequency of charging piles A and charging piles B is similar during peak hours, their correlation is recorded and the common characteristics of their power changes are analyzed. According to the mined power linkage law, the associated charging piles close to the specific charging pile are identified. These charging piles usually have similar geographical locations or power fluctuations. Using geographic information system (GIS) data, the distance between charging piles is determined, and charging piles with a distance of less than 500 meters are marked as adjacent charging piles. Analyze the identified adjacent charging piles and record their power characteristics and coupling relationships for subsequent analysis and decision-making. Generate a table containing the adjacent charging pile ID, power characteristics, and coupling relationships to help identify potential collaborative charging opportunities. Based on the identified adjacent associated charging piles, use graph theory to analyze the power collaborative path between charging piles. Ensure that the path analysis can reflect the power flow relationship between each charging pile. Use the Dijkstra algorithm to calculate the shortest path for each charging pile and identify the best path for power flow.

[0035] The power coordination paths of adjacent charging piles are analyzed to extract the power coordination path characteristics of each charging pile to ensure that the data can reflect the dynamic characteristics of power flow. The power flow path from charging pile A to charging pile B is recorded to ensure that the path information is detailed and accurate. The conflict detection algorithm is used to analyze the power conflict situation between charging piles and identify the power conflict situation that may occur during the peak charging period. The conflict detection threshold is set to the maximum power limit. If multiple charging piles request to exceed this limit at the same time, it is recorded as a conflict state. The conflict situation analysis of the power flow of the identified charging piles is performed to evaluate the power request situation of each charging pile during the peak charging period and identify potential conflicts. If charging pile A requests 50 kW of power at a certain moment, charging pile B also requests 40 kW, and the total power limit of the system is 80 kW, it is recorded as a conflict situation. Use time series prediction models (such as ARIMA models or long short-term memory networks LSTM) to predict the power flow between charging piles. Selecting a suitable model can improve the accuracy of the prediction. The LSTM model is selected because it can handle long-term dependencies in time series data and is suitable for analyzing the dynamic changes of charging power. Use historical charging pile power data to train the selected prediction model to ensure that the model can learn the power flow pattern between charging piles. Use the power data from the past 30 days for training to ensure that the model can capture the regularity and changes of charging peaks.

[0036] The trained model is used to predict the future power flow and identify the power flow direction and intensity of each charging pile. It is predicted that in the next hour, the power flow of charging pile A will be 60 kW and the power flow of charging pile B will be 40 kW. Ensure that the data is accurate and record the timestamp. Select an appropriate topology evolution method, usually dynamic network analysis, to ensure that the changes in the power coordination path between charging piles can be reflected. The network reconstruction algorithm is used to dynamically adjust the connection relationship between charging piles based on the predicted power flow and historical data. The power coordination path of each charging pile is topologically evolved according to the predicted power flow to ensure that the path can adapt to environmental changes and usage needs. If the power request of a charging pile increases significantly during peak hours, its coordination path with other charging piles is dynamically adjusted to ensure the stability and efficiency of the system. The topology evolution results are recorded in the database and a report is formed for subsequent analysis and decision-making. A topology diagram of the coordination path of charging piles is generated to show the connection relationship and power flow path between charging piles to help managers optimize and adjust.

[0037] In this embodiment, refer to Figure 4 , is a flowchart of detailed implementation steps of step S3. In this embodiment, the detailed implementation steps of step S3 include: Identify the idle state of the charging pile according to the real-time operation monitoring parameters of the charging pile, and extract the idle available charging piles and the occupied charging piles; Analyze the time-series charging power changes of the occupied charging piles and generate the time-series charging power change characteristics; Calculating the number of idle available charging piles and performing idle time distribution prediction to obtain an idle charging pile time distribution prediction value; According to the time distribution prediction value of idle charging piles and the time sequence charging power change characteristics, the charging pile power trend is predicted, and the charging pile power trend prediction status in the future time period is generated; The power peak congestion evolution of the power trend prediction state of the charging pile is performed according to the power cooperative path network to generate the power peak congestion state characteristics of the charging pile.

[0038] In this embodiment, the current operating parameters are extracted from the real-time monitoring system of the charging pile, including the charging status (idle or occupied), charging start time and end time of each charging pile. Ensure the timeliness and accuracy of the data for subsequent analysis. Set the charging pile status data to be obtained from the system every 30 seconds, and record the status of each charging pile (such as "idle" and "charging"). Classify the collected charging pile status data, and divide the charging piles into "idle and available charging piles" and "occupied charging piles". Mark the corresponding attributes of the charging piles according to their current status. If a charging pile is displayed as "charging" at the current moment, it is marked as "occupied"; otherwise, it is marked as "idle and available". Record the status information of each charging pile, including the timestamps of idle and occupied. Extract the relevant data of the occupied charging piles from the charging pile status records, mainly including charging power, charging time and state change time, which will provide basic data for charging power change analysis. Record the power change information of each occupied charging pile during the charging process to ensure that the data covers the integrity of the charging cycle. Perform time series analysis on the charging power of the occupied charging piles and calculate the power change characteristics. Statistical methods (such as moving average, variance, etc.) can be used to analyze the trend of power changes over time. Set a time window (such as every 5 minutes), calculate the average charging power of each occupied charging pile within the window, and record its changes. Calculate the number of currently idle and available charging piles based on the charging pile status records. Ensure that the data is accurate for subsequent idle time distribution analysis. Count the number of charging piles with the current status of "idle". If there are currently 10 charging piles in an idle state, record the number. For the time distribution forecast of idle and available charging piles, time series analysis methods such as ARIMA model or exponential smoothing method can be used to predict the future idle time distribution. Use the idle time data of the past two weeks for analysis to predict the idle time distribution of charging piles in the next 24 hours and calculate the expected idle time. Record the forecast results in the database and generate a forecast report on the time distribution of idle charging piles for subsequent management and decision-making. Generate a forecast report to record the expected number of idle charging piles in each time period to help managers allocate and optimize charging piles. According to the time series charging power change characteristics of the occupied charging piles, select a suitable power trend prediction model (such as linear regression, LSTM, etc.) to predict the power trend of the charging piles. Select the LSTM model because it is suitable for processing time series data and can effectively capture the time dependence of power changes. Use historical charging power data to train the selected prediction model to ensure that the model can learn the law of power changes. Use training sets and validation sets to evaluate the model to ensure accuracy. Use the charging power data of the past 30 days for training to verify the prediction ability of the model and ensure that the error is within an acceptable range. Apply the trained model to predict the trend of the charging pile power in the future time period and generate the expected power change status.Predict the charging power changes in the next 6 hours and record the predicted power values ​​for each time period for subsequent analysis and optimization. According to the power trend prediction status of the charging piles, analyze the power coordination path network to identify potential power peak congestion. Ensure that the network analysis can reflect the mutual influence between the charging piles. Use graph theory to construct a power coordination path network between charging piles, with nodes representing charging piles and edges representing power flow relationships. Based on the analysis results of the power coordination path network, identify the power request of the charging piles during peak periods and mark possible congestion states. If it is found that multiple charging piles request power in the same time period that exceeds the maximum carrying capacity of the system, it will be recorded as a power peak congestion state, and its occurrence time and impact range will be marked.

[0039] In this embodiment, step S4 includes the following steps: Acquire new energy vehicle flow data of the service area; divide the new energy vehicle flow data into time windows, and extract vehicle flow data of multiple time windows; The average vehicle flow per unit time is calculated for the vehicle flow data of multiple time windows, and the average vehicle flow per unit time of each time window is extracted; Performing dynamic traffic flow differentiation analysis based on the average traffic flow per unit time, and extracting traffic flow differentiation features of multiple time windows; Identify the current cycle situation based on the new energy vehicle flow data in the service area; Obtain the service area's historical traffic flow record log; match the service area's historical traffic flow record log with the corresponding period according to the current time period, and analyze the traffic flow changes to generate periodic change rules; The traffic flow is predicted based on the differentiated characteristics of the traffic flow according to the periodic variation law to obtain a traffic flow prediction heat map.

[0040] In this embodiment, the source of new energy vehicle flow data is determined, which usually includes the traffic monitoring system, sensors, cameras and traffic flow statistics equipment in the service area. Ensure the integrity and accuracy of the data for subsequent analysis. Set the traffic flow data to be obtained from the traffic monitoring system every 5 minutes, and record the number of new energy vehicles entering the service area in each time period. By writing a data acquisition program or using an existing API interface, the new energy vehicle flow data is obtained regularly and stored in the database. Ensure that the data format is unified for subsequent processing. Create a database table with storage fields including "timestamp", "traffic flow", etc. to ensure that this information can be recorded for each data collection. Verify the acquired traffic flow data to ensure that the data is complete and free of abnormal values. Remove duplicate records and unreasonable abnormal values ​​(such as negative values ​​or traffic flow exceeding expectations). Set the reasonable range of traffic flow to 0 to 1000 vehicles, and any value outside this range should be marked as abnormal and recorded. According to the analysis requirements, divide the time into multiple windows (such as every hour or every half hour) to facilitate segmented analysis of traffic flow data. Ensure that the data in each time window is complete. Set every 30 minutes as a time window and record the traffic flow data in each time period. Sort the data in the divided time windows and extract the traffic flow data of each time window. Ensure that the data in each window can reflect the traffic flow situation in the time period. Count the total traffic flow in each time window and record the start and end time of each window. Calculate the average per unit time of the traffic flow data in each time window. The traffic flow per unit time is obtained by dividing the traffic flow in each window by the window length. For a 30-minute time window, if the traffic flow in the window is 300 vehicles, the traffic flow per unit time is 300 vehicles / 0.5 hours = 600 vehicles / hour. Record the calculated average traffic flow per unit time in the database to form a traffic flow statistics report for subsequent analysis. Generate a table containing "time window" and "average traffic flow per unit time" to record the average traffic flow situation in each time window. Select a suitable differentiation analysis method, such as t-test or variance analysis, to compare the average traffic flow per unit time of different time windows and extract the differentiated characteristics of traffic flow. Set the significance level to 0.05 to ensure that statistically significant differences in traffic volume can be identified. Perform differential analysis on the average traffic volume per unit time of multiple time windows to identify time periods with significant changes in traffic volume. Record the analysis results and mark the time windows with significant differences. If significant differences in traffic volume are found between the morning peak and the evening peak, record the relevant information of these time windows. Use periodic analysis methods, such as Fourier transform or autocorrelation analysis, to identify the periodicity of traffic volume at the current time to ensure that the periodic characteristics of traffic volume can be captured. Use autocorrelation functions to analyze traffic volume data and identify the frequency of periodic fluctuations. Perform periodic analysis on the traffic volume data at the current time to identify the periodic characteristics of traffic volume.Record the analysis results and mark the periodic features. If the analysis results show that the traffic volume has a 24-hour periodic fluctuation, record the periodic features and related parameters. Extract the historical traffic volume record logs of the service area from the database to ensure that the data covers a long enough period of time for periodic change analysis. Extract the traffic volume records of the past 6 months to ensure the integrity and accuracy of the data. According to the current periodic situation, match the corresponding period of the service area historical traffic volume record logs and analyze the relationship between the historical data and the current period. If the current identified period is 24 hours, extract the corresponding 24-hour data from the historical records for analysis and compare the changes in traffic volume. Select a suitable prediction model, such as linear regression, time series prediction model (such as ARIMA or LSTM), and predict traffic volume according to the periodic change law to obtain the future traffic volume state. Select the ARIMA model for prediction to ensure that the trend and seasonal changes of historical data can be captured. Use historical traffic volume data to train the selected prediction model to ensure that the model can learn the change law of traffic volume. Use the training set and validation set to evaluate the accuracy of the model. Use the traffic volume data of the past 3 months for training and validation to ensure that the prediction error is within an acceptable range. Use the trained model to predict traffic flow in future time periods, generate traffic flow prediction values, and record the prediction results for each time period. Predict traffic flow changes in the next 24 hours and record the prediction values ​​for each time period. Generate a traffic flow heat map based on the prediction results to visualize the distribution of traffic flow in different time periods, helping managers quickly identify peak and trough periods. Use the heat map tool to display traffic flow changes in the next 24 hours, using different colors to represent different traffic flow levels for easier decision-making.

[0041] In this embodiment, the specific steps of step S5 are: Identify the user charging queue according to the real-time operation monitoring parameters of the charging pile; The queuing time of the user charging queue is counted, and the total length of time in and out of the queue is calculated to obtain the total length of the queue; Perform charging demand analysis based on the total queue length and traffic flow prediction heat map to generate user charging demand characteristics; Based on the power peak congestion state characteristics, a comprehensive charging power demand prediction is performed on the user's charging demand characteristics to generate comprehensive charging power demand prediction data; Perform intelligent advance charging power scheduling based on comprehensive charging power demand forecast data and build an intelligent charging power scheduling strategy.

[0042] In this embodiment, the user charging status data is extracted from the real-time monitoring system of the charging pile, including the current usage status of each charging pile (such as idle, charging, queuing, etc.). Ensure the accuracy and timeliness of the data so as to identify the charging queue of the user. Set the charging pile status information to be collected every 5 minutes, and record the queue status, current number of users and queuing time of each charging pile. Adopt the queue identification method based on status monitoring, analyze the usage status of each charging pile, and identify the users who are queuing. It is possible to determine whether the user is in the queue state by setting a threshold (such as the charging pile usage status is "queuing" and there are users waiting). If the status of a charging pile is "charging" and there are users waiting later, these users are recorded as queuing users, and a queue list is created. The identified user charging queue data is recorded in the database, including user ID, queue time, charging pile ID and other information, to provide basic data for subsequent analysis. Generate a table containing "user ID", "charging pile ID" and "queue time" to ensure the structured and queryable data. Statistics are performed on the queuing time of the user charging queue, and the queuing time of each user is calculated. The queue time can be calculated by recording the time when the user enters the queue and the time when charging starts. If a user enters the queue at 08:00 and starts charging at 08:15, the queue time is 15 minutes. Count the queue time of all users and calculate the total queue time. That is, the sum of the queue time of all users, and record it for subsequent analysis. If there are 5 users with queue times of 15 minutes, 10 minutes, 20 minutes, 5 minutes and 30 minutes in a certain period, the total queue time is 15 + 10 + 20 + 5 + 30 = 80 minutes. According to the total queue time and the traffic flow prediction heat map, use demand analysis methods such as regression analysis or time series analysis to analyze the user's charging demand characteristics. Ensure that the changing trend of demand can be identified. Combined with the peak period data in the traffic flow prediction heat map, analyze the charging demand characteristics of users during peak periods. Comprehensively analyze the counted queue time and traffic flow data to extract the user's charging demand characteristics, including peak demand and duration. If the total queue length is 80 minutes and the vehicle flow is 200 vehicles during a peak period, the charging demand characteristics of this period are recorded and a demand analysis report is generated. Based on the previously identified power peak congestion state characteristics, the relationship between the charging demand of users during peak hours and the power availability of charging piles is analyzed for comprehensive prediction. If the power peak congestion situation of a certain period is recorded, the gap between the charging demand and the power supply capacity of this period is analyzed. Select a suitable prediction model, such as a multivariate linear regression model or a machine learning algorithm (such as random forest or LSTM), to make a comprehensive prediction of user charging demand to generate comprehensive charging power demand prediction data. Use historical charging demand data and current charging demand characteristics for model training to ensure that the model can capture the complex pattern of charging demand.Use the trained model to predict future charging demand and generate comprehensive charging power demand forecast data to ensure that the data can reflect future demand changes. Predict charging demand within the next hour and record the predicted power demand value for each time period. Based on the comprehensive charging power demand forecast data, design an intelligent scheduling strategy to optimize the power allocation of charging piles to ensure that charging resources can be reasonably scheduled during high demand periods. Set charging power to be allocated to users with longer queues during peak charging demand periods. Implement the designed charging power intelligent scheduling strategy and monitor it to verify its effectiveness. Monitor the usage of charging piles and user queues in real time to ensure that the scheduling strategy can effectively improve charging efficiency. Monitor the changes in user queue time after implementation. If the queue time is reduced, it means that the scheduling strategy is effective.

[0043] In this embodiment, the specific steps of step S6 are: Perform abnormal voltage fluctuation analysis on the multi-dimensional feature matrix and extract abnormal voltage fluctuation data; Identify abnormal current thermal drift, abnormal plug-in and plug-out frequency, and thermal runaway trigger records based on a multi-dimensional feature matrix; Predict charging pile failures based on abnormal voltage fluctuation data, abnormal current thermal drift, abnormal plugging and unplugging frequency, and thermal runaway trigger records, and mark faulty charging piles; According to the faulty charging pile, the charging power intelligent scheduling strategy is used to shield the faulty charging pile, make global power rescheduling decisions, and build an intelligent charging management engine.

[0044] In this embodiment, voltage data is extracted from the previously generated multi-dimensional feature matrix to ensure that the data includes the voltage readings of each charging pile in different time periods. Ensure the timeliness and accuracy of the data for abnormal analysis. Set the data acquisition cycle to every second, obtain voltage data from the charging pile, and record the voltage value of each charging pile at a specific time point. Use statistical analysis methods (such as Z-score analysis or IQR method) to detect whether the voltage fluctuation is abnormal. Set a reasonable threshold to identify the situation where the voltage fluctuation exceeds the normal range. Set the Z-score threshold to ±3. If the Z-score of the voltage reading of a charging pile exceeds the threshold, it is marked as abnormal. The detected abnormal voltage data is extracted and recorded, including the time when the abnormality occurred, the charging pile ID and the abnormal voltage value, which will be used for subsequent fault prediction. If a charging pile records a voltage of 450V at 08:15 (the normal range is 100V-400V), the time and abnormal voltage value are recorded. Extract current data from the multi-dimensional feature matrix and analyze the current fluctuation of the charging pile. Ensure the integrity of the data for thermal drift detection. Set the normal current range to 0A-100A. If the current fluctuation exceeds this range for multiple consecutive times, it is marked as abnormal thermal drift. Count the plug-in and unplug frequency of the charging piles and record the number of plug-ins and unplugs of each charging pile within a certain period of time. High-frequency plug-ins and unplugs may cause equipment damage or failure. If a charging pile is plugged and unplugged more than 5 times within an hour, record the ID and plug-in number of the charging pile and mark it as abnormal. Monitor the temperature data of the charging pile to identify whether there is a trigger record for thermal runaway risk. Set a temperature threshold (such as 80°C). If the threshold is reached, record the event. If the temperature of a charging pile exceeds 80°C during charging, record the event and associate it with other abnormal data. Select a suitable fault prediction model, such as decision tree, random forest, or support vector machine (SVM), to comprehensively analyze multiple features such as abnormal voltage, abnormal current thermal drift, and plug-in frequency. Select the random forest model because it can process multi-dimensional data and has good classification performance. Use historical fault data to train the selected model to ensure that the model can learn the difference between fault characteristics and normal conditions. Divide the data into training set and test set for model validation. Use the fault data of the past year for training, verify the accuracy and recall of the model to ensure its effectiveness in fault detection. Apply the trained model to predict the fault status of the current charging pile and mark the charging piles with potential faults. Record the prediction results for each charging pile. If the model predicts that the charging pile ID "C001" has a fault risk, record the charging pile and its failure probability. Identify and record the charging piles marked as faulty to ensure that they can be blocked in time in the scheduling decision. Avoid assigning faulty charging piles to users. Generate a list of faulty charging piles, including the charging pile ID, fault type, and predicted probability.According to the identification results of the faulty charging pile, the existing intelligent charging power scheduling strategy is adjusted to ensure that the remaining charging resources can still be effectively allocated when the faulty charging pile is shielded. If a charging pile fails during peak hours, its load is allocated to other idle charging piles to ensure that the overall charging efficiency is not affected. According to the shielding processing results of the faulty charging pile, a global power rescheduling decision is made. Ensure that the system can still operate efficiently in the event of a fault. Calculate the load capacity and expected charging demand of each available charging pile, and reallocate power to minimize user waiting time. Design the architecture of the intelligent charging management engine, including the data input module (receive real-time monitoring data), the fault detection module, the scheduling decision module, and the data output module (generate scheduling results). Ensure that the engine can process the monitored data in real time and can quickly make fault detection and scheduling decisions. Integrate the various modules to ensure smooth data flow, and perform functional tests with actual data to verify the effectiveness and stability of the engine. Use simulated data for testing to observe the engine's response ability under different fault conditions to ensure that it can quickly adjust the scheduling strategy. During the operation of the engine, monitor the scheduling effect, collect user feedback and system performance data, and perform continuous optimization. Record users’ charging waiting time and charging success rate, analyze the effectiveness of scheduling strategies, and make adjustments based on feedback.

[0045] In this embodiment, a service area new energy charging management system based on multi-device data analysis is provided, which is used to execute the service area new energy charging management method based on multi-device data analysis as described above, including: The multi-dimensional feature perception module is used to obtain the real-time operation monitoring parameters of all charging piles in the service area; and to perform multi-dimensional feature perception and charging behavior time heat distribution analysis to build a dynamic charging behavior hotspot map; The power collaborative path module is used to perform multi-time point power sampling according to the dynamic charging behavior hot spot map, and to perform power collaborative path topology evolution to build a power collaborative path network; The power trend prediction module is used to predict the power trend of charging piles and the evolution of power peak congestion according to the power cooperative path network, and generate the power peak congestion state characteristics of the charging piles; The vehicle flow prediction module is used to obtain the new energy vehicle flow data in the service area, calculate the average vehicle flow per unit time, and predict the vehicle flow to obtain a vehicle flow prediction heat map; A power intelligent dispatching module, used to perform comprehensive charging power demand prediction and pre-charging power intelligent dispatching based on the power peak congestion state characteristics according to the traffic flow prediction heat map, and build a charging power intelligent dispatching strategy; The fault shielding module is used to shield faulty charging piles for the charging power intelligent scheduling strategy, make global power rescheduling decisions, and build an intelligent charging management engine.

[0046] The present invention can timely reflect the working status and usage of charging piles by acquiring real-time operation monitoring parameters of all charging piles, which provides an accurate data basis for subsequent analysis and decision-making. Multi-dimensional feature perception and time heat distribution analysis of charging behavior can identify the peak time and hot spots of charging behavior, which helps to understand the charging habits and needs of users and optimize the configuration of charging resources. A dynamic charging behavior hot spot map is constructed so that managers can intuitively understand the usage of charging piles and support decision-making and resource allocation. Multi-point power sampling is performed according to the dynamic charging behavior hot spot map, which can capture the power changes in different time periods and provide a basis for power management. The power collaborative path topology evolution is performed to construct a power collaborative path network, which can help identify the power relationship between charging piles and support more efficient power scheduling and management. The power trend prediction of charging piles through the power collaborative path network can identify the changes in power demand in advance and help managers effectively deal with potential power peak congestion problems. The power peak congestion state characteristics of the charging piles are generated, which enables managers to quickly identify and take measures to avoid the decline in user experience due to insufficient power. Obtaining the traffic data of new energy vehicles in the service area and calculating the average traffic volume per unit time can provide support for charging demand prediction. Through traffic flow prediction, a traffic flow prediction heat map is generated, so that managers can intuitively understand the changes in traffic flow in different time periods, thereby optimizing the distribution and scheduling of charging piles. According to the traffic flow prediction heat map, a comprehensive charging power demand prediction is made for the power peak congestion state characteristics, which helps to reasonably allocate charging resources and reduce peak loads. Constructing a charging power intelligent scheduling strategy can reasonably schedule charging power in different time periods, improve the utilization rate of charging piles, and reduce user waiting time. The charging power intelligent scheduling strategy shields faulty charging piles to ensure that when a charging pile fails, the system can automatically adjust the power allocation to minimize user inconvenience. Make global power rescheduling decisions and build an intelligent charging management engine so that the entire system can quickly adapt to dynamic environments and ensure the stability and continuity of charging services.

[0047] Therefore, the embodiments should be regarded as illustrative and non-restrictive from all points, and the scope of the present invention is limited by the appended claims rather than the above description, and it is therefore intended that all changes falling within the meaning and range of equivalent elements of the application documents are included in the present invention.

[0048] The above is only a specific embodiment of the present invention, so that those skilled in the art can understand or implement the present invention. Various modifications to these embodiments will be apparent to those skilled in the art, and the general principles defined herein can be implemented in other embodiments without departing from the spirit or scope of the present invention. Therefore, the present invention will not be limited to the embodiments shown herein, but should conform to the widest scope consistent with the principles and novel features invented herein.

Claims

1. A service area new energy charging management method based on multi-device data analysis, characterized in that: The following steps are involved: Step S1: Obtain real-time operation monitoring parameters of all charging piles in the service area; It also conducts multi-dimensional feature perception and charging behavior time heat distribution analysis to build a dynamic charging behavior hotspot map; Step S2: Perform multi-time point power sampling according to the dynamic charging behavior hotspot map, and perform power coordination path topology evolution to build a power coordination path network; Step S3: predicting the power trend of the charging pile and the evolution of the power peak congestion according to the power cooperative path network, and generating the power peak congestion state characteristics of the charging pile; Step S4: acquiring new energy vehicle flow data of the service area, calculating the average vehicle flow per unit time, and performing vehicle flow prediction to obtain a vehicle flow prediction heat map; Step S5: performing comprehensive charging power demand prediction and pre-charging power intelligent scheduling based on the power peak congestion state characteristics according to the vehicle flow prediction heat map, and constructing a charging power intelligent scheduling strategy; Step S6: Shield the faulty charging piles in the charging power intelligent scheduling strategy, make global power rescheduling decisions, and build an intelligent charging management engine.

2. The service area new energy charging management method based on multi-device data analysis according to claim 1 is characterized in that: The specific steps of step S1 are: Obtaining real-time operation monitoring parameters of all charging piles in the service area; calculating the voltage and current parameters of the charging piles according to the real-time operation monitoring parameters of the charging piles to obtain the electrical characteristics of each charging pile; The usage frequency of the real-time operation monitoring parameters of the charging pile is successively counted to extract the usage frequency parameters of the charging pile; Extracting the charging time and idle time period of the real-time operation monitoring parameters of the charging pile; Perform multi-dimensional feature perception on the usage frequency parameters of the electrical characteristic charging pile, the charging time and the idle period, and construct a multi-dimensional feature matrix; Perform time stamp synchronization on the multi-dimensional feature matrix, perform time vectorization decomposition, and construct a rasterized time window for charging behavior; Each charging behavior is extracted according to the rasterized time window of charging behavior, and the time heat distribution analysis of charging behavior is performed to construct a dynamic charging behavior hotspot map.

3. The service area new energy charging management method based on multi-device data analysis according to claim 1 is characterized in that: The specific steps of step S2 are: Calculate the power of each charging behavior based on the dynamic charging behavior hotspot map to generate the power characteristics of each charging pile; Performing power sampling at multiple time points on the charging power characteristics to construct charging power curves for multiple charging piles; Perform discrete Fourier transform on the charging power curves of multiple charging piles to generate charging power frequency domain data; Perform spectrum peak detection on the charging power frequency domain data and mark multiple spectrum peaks; Identify the peak timing shift characteristics of the spectrum peak and mark the high-frequency charge and discharge fluctuation nodes; The power collaborative path topology evolves according to the high-frequency charging and discharging fluctuation nodes, and constructs a power collaborative path network.

4. The service area new energy charging management method based on multi-device data analysis according to claim 3 is characterized in that: The specific steps of performing power cooperative path topology evolution according to high-frequency charging and discharging fluctuation nodes and constructing a power cooperative path network are as follows: Multi-cycle trend encoding is performed on high-frequency charging and discharging fluctuation nodes to generate charging power fluctuation change chains for different charging piles; Perform power fluctuation coupling detection on the charging power fluctuation chain, and extract and identify the power fluctuation coupling relationship between charging piles; The power fluctuation coupling relationship is mined to analyze the power correlation changes of multiple charging piles to generate the power linkage law between charging piles; Identify nearby associated charging piles based on the charging power fluctuation chain; According to the power linkage law between charging piles, the potential power coordination path of the adjacent associated charging piles is analyzed to extract the power coordination path of each charging pile; Analyze the power conflict situation of the power linkage between charging piles to obtain the power conflict situation between charging piles; Predict power flow based on power conflict between charging piles; Based on the power flow direction, the power cooperative path of each charging pile is topologically evolved to construct a power cooperative path network.

5. The service area new energy charging management method based on multi-device data analysis according to claim 1 is characterized in that: The specific steps of step S3 are: Identify the idle state of the charging pile according to the real-time operation monitoring parameters of the charging pile, and extract the idle available charging piles and the occupied charging piles; Analyze the time-series charging power changes of the occupied charging piles and generate the time-series charging power change characteristics; Calculating the number of idle available charging piles and performing idle time distribution prediction to obtain an idle charging pile time distribution prediction value; According to the time distribution prediction value of idle charging piles and the time sequence charging power change characteristics, the charging pile power trend is predicted, and the charging pile power trend prediction status in the future time period is generated; The power peak congestion evolution of the power trend prediction state of the charging pile is performed according to the power cooperative path network to generate the power peak congestion state characteristics of the charging pile.

6. The service area new energy charging management method based on multi-device data analysis according to claim 1 is characterized in that: The specific steps of step S4 are: Acquire new energy vehicle flow data of the service area; divide the new energy vehicle flow data into time windows, and extract vehicle flow data of multiple time windows; The average vehicle flow per unit time is calculated for the vehicle flow data of multiple time windows, and the average vehicle flow per unit time of each time window is extracted; Performing dynamic traffic flow differentiation analysis based on the average traffic flow per unit time, and extracting traffic flow differentiation features of multiple time windows; Identify the current cycle situation based on the new energy vehicle flow data in the service area; Obtain the service area's historical traffic flow record log; match the service area's historical traffic flow record log with the corresponding period according to the current time period, and analyze the traffic flow changes to generate periodic change rules; The traffic flow is predicted based on the differentiated characteristics of the traffic flow according to the periodic variation law to obtain a traffic flow prediction heat map.

7. The service area new energy charging management method based on multi-device data analysis according to claim 1 is characterized in that: The specific steps of step S5 are: Identify the user charging queue according to the real-time operation monitoring parameters of the charging pile; The queuing time of the user charging queue is counted, and the total length of time in and out of the queue is calculated to obtain the total length of the queue; Perform charging demand analysis based on the total queue length and traffic flow prediction heat map to generate user charging demand characteristics; Based on the power peak congestion state characteristics, a comprehensive charging power demand prediction is performed on the user's charging demand characteristics to generate comprehensive charging power demand prediction data; Perform intelligent advance charging power scheduling based on comprehensive charging power demand forecast data and build an intelligent charging power scheduling strategy.

8. The service area new energy charging management method based on multi-device data analysis according to claim 1 is characterized in that: The specific steps of step S6 are: Perform abnormal voltage fluctuation analysis on the multi-dimensional feature matrix and extract abnormal voltage fluctuation data; Identify abnormal current thermal drift, abnormal plug-in and plug-out frequency, and thermal runaway trigger records based on a multi-dimensional feature matrix; Predict charging pile failures based on abnormal voltage fluctuation data, abnormal current thermal drift, abnormal plugging and unplugging frequency, and thermal runaway trigger records, and mark faulty charging piles; According to the faulty charging pile, the charging power intelligent scheduling strategy is used to shield the faulty charging pile, make global power rescheduling decisions, and build an intelligent charging management engine.

9. A service area new energy charging management system based on multi-device data analysis, characterized in that: The method for managing new energy charging in a service area based on multi-device data analysis as claimed in claim 1 comprises: The multi-dimensional feature perception module is used to obtain the real-time operation monitoring parameters of all charging piles in the service area; and to perform multi-dimensional feature perception and charging behavior time heat distribution analysis to build a dynamic charging behavior hotspot map; The power collaborative path module is used to perform multi-time point power sampling according to the dynamic charging behavior hot spot map, and to perform power collaborative path topology evolution to build a power collaborative path network; The power trend prediction module is used to predict the power trend of charging piles and the evolution of power peak congestion according to the power cooperative path network, and generate the power peak congestion state characteristics of the charging piles; The vehicle flow prediction module is used to obtain the new energy vehicle flow data in the service area, calculate the average vehicle flow per unit time, and predict the vehicle flow to obtain a vehicle flow prediction heat map; A power intelligent dispatching module, used to perform comprehensive charging power demand prediction and pre-charging power intelligent dispatching based on the power peak congestion state characteristics according to the traffic flow prediction heat map, and build a charging power intelligent dispatching strategy; The fault shielding module is used to shield faulty charging piles for the charging power intelligent scheduling strategy, make global power rescheduling decisions, and build an intelligent charging management engine.

Images