Charging station power control method and device combined with long-term monitoring and analysis

Abstract

The present invention discloses a power control method and device for a charging station combined with long-term monitoring and analysis, and relates to the field of charging management technology. The method comprises: for a charging station, collecting equipment configuration and architecture, and constructing a station diagram structure. At the same time, a timing monitoring and analysis channel is established, including long-term and short-term monitoring, to evaluate the power conversion state. Based on the station diagram, equipment-independent and station-global analysis links are established, and the timing monitoring channel is accessed to obtain a power conversion evaluation. Through analysis and evaluation, the abnormal operation risks of charging equipment and stations are identified. Finally, for abnormal risks, the control targets are located, control analysis is performed, and power control strategies are formulated to smooth or eliminate risks. Thereby achieving the technical effect of improving load adaptability and simplifying control strategies.

Description

Technical Field

[0001] The present invention relates to the field of charging management technology, and in particular to a charging station power control method and device combined with long-term time-series monitoring and analysis. Background Art

[0002] With the popularity of electric vehicles, charging stations, as the core infrastructure for electric vehicle power supply, bear huge power supply pressure. Power control of charging stations has become a key link to ensure their safe, stable and efficient operation. During the operation of charging stations, the charging load is unbalanced due to the randomness of electric vehicle charging and the fluctuation of charging demand during peak hours. The existing power control methods are mostly focused on short-term predictions, lacking the mining and analysis of long-term series data, resulting in insufficient grasp of long-term load fluctuation trends. There are technical problems such as poor load adaptability and complex control strategies. Summary of the invention

[0003] The present invention provides a charging station power control method and device combined with long-term monitoring and analysis to solve the technical problems of poor load adaptability and complex control strategies in the prior art, and achieve the technical effects of improving load adaptability and simplifying control strategies.

[0004] In a first aspect, the present invention provides a charging station power control method combined with long-term monitoring and analysis, wherein the method comprises:

[0005] Collect the charging equipment configuration and architecture of the target charging station and establish the station map structure.

[0006] A timing monitoring and analysis channel is established, and the timing monitoring and analysis channel is used to evaluate the power conversion status of the charging station. The timing monitoring and analysis channel includes a long timing monitoring and analysis channel and a short timing monitoring and analysis channel.

[0007] Based on the site diagram structure, an equipment-independent analysis link and a site-global analysis link are established, and the equipment-independent analysis link and the site-global analysis link are respectively connected to the timing monitoring and analysis channel to obtain equipment-independent long-time / short-time power conversion evaluation and site-global long-time / short-time power conversion evaluation.

[0008] According to the independent long-time / short-time power conversion evaluation of the equipment and the global long-time / short-time power conversion evaluation of the station, the operating power of the charging equipment and the station is analyzed to identify abnormal risks.

[0009] The control target is located according to the abnormal risk, and the control target is analyzed based on the abnormal risk to obtain a power control strategy, where the power control strategy is used to smooth or eliminate the abnormal risk.

[0010] In a second aspect, the present invention further provides a charging station power control device combined with long-term monitoring and analysis, wherein the device comprises:

[0011] A graph structure building module is used to collect the charging equipment configuration and architecture of the target charging station and establish a station graph structure.

[0012] A timing analysis construction module, wherein the timing analysis construction module is used to establish a timing monitoring and analysis channel, wherein the timing monitoring and analysis channel is used to evaluate the power conversion status of the charging station, and the timing monitoring and analysis channel includes a long timing monitoring and analysis channel and a short timing monitoring and analysis channel.

[0013] An analysis link establishment module is used to establish a device-independent analysis link and a site-global analysis link based on the site map structure, and connect the device-independent analysis link and the site-global analysis link to the timing monitoring and analysis channel respectively to obtain a device-independent long-time / short-time power conversion evaluation and a site-global long-time / short-time power conversion evaluation.

[0014] The abnormality identification module is used to analyze the operating power of the charging equipment and the station according to the independent long-time / short-time power conversion evaluation of the equipment and the global long-time / short-time power conversion evaluation of the station, and identify abnormal risks.

[0015] An analysis and control module is used to locate a control target according to the abnormal risk, perform control analysis on the control target based on the abnormal risk, and obtain a power control strategy, wherein the power control strategy is used to smooth or eliminate the abnormal risk.

[0016] The present invention discloses a charging station power control method and device combined with long-term monitoring and analysis, including: collecting the charging equipment configuration and architecture information of the target charging station, and establishing a graph structure model of the station; constructing a timing monitoring and analysis channel to monitor and evaluate the power conversion state of the charging station, wherein the timing monitoring and analysis channel includes a long-term monitoring channel and a short-term monitoring channel; based on the station graph structure, establishing a device-independent analysis link and a station-global analysis link, respectively, and connecting these links to the timing monitoring and analysis channel, so as to obtain a device-independent long-term / short-term power conversion evaluation and a station-global long-term / short-term power conversion evaluation; based on the device-independent and station-global long-term / short-term power conversion evaluation, analyzing the overall operating power situation of the charging equipment and the station, and identifying potential abnormal risks; locating the target equipment or area according to the identified abnormal risks, performing regulation and analysis based on the risk characteristics, and formulating a power control strategy for smoothing power fluctuations or eliminating potential abnormal risks. The charging station power control method and device combined with long-term monitoring and analysis disclosed in the present invention solve the technical problems of poor load adaptability and complex control strategies, and achieve the technical effects of improving load adaptability and simplifying control strategies. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] Figure 1 It is a flow chart of a charging station power control method combined with long-term monitoring and analysis of the present invention;

[0018] Figure 2 It is a structural schematic diagram of a charging station power control device combined with long-term monitoring and analysis of the present invention.

[0019] Explanation of the accompanying drawings: graph structure construction module 11, timing analysis construction module 12, analysis link establishment module 13, anomaly identification module 14, analysis and regulation module 15. DETAILED DESCRIPTION

[0020] The technical solution provided in the embodiments of the present invention is to solve the technical problems of poor load adaptability and complex control strategies in the prior art. The overall idea adopted is as follows:

[0021] First, the charging equipment configuration and architecture of the target charging station are collected to establish a station map structure; then, a timing monitoring and analysis channel is established, which is used to evaluate the power conversion state of the charging station and includes a long-time monitoring and analysis channel and a short-time monitoring and analysis channel; then, based on the station map structure, an independent equipment analysis link and a station global analysis link are established, and the independent equipment analysis link and the station global analysis link are respectively connected to the timing monitoring and analysis channel to obtain an independent long-time / short-time power conversion evaluation of the equipment and a global long-time / short-time power conversion evaluation of the station; then, according to the independent long-time / short-time power conversion evaluation of the equipment and the global long-time / short-time power conversion evaluation of the station, the operating power of the charging equipment and the station is analyzed to identify abnormal risks; then, according to the abnormal risks, the control target is located, and the control target is analyzed based on the abnormal risks to obtain a power control strategy; finally, the power control strategy is used to control the charging equipment and the station to smooth or eliminate abnormal risks.

[0022] The above technical solution will be described in detail below in conjunction with the accompanying drawings and specific implementation methods of the specification to better understand the above technical solution. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than all of the embodiments of the present invention. It should be understood that the present invention is not limited to the example embodiments used only to explain the present invention. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention. In addition, it should be noted that, for the convenience of description, only the parts related to the present invention are shown in the drawings, rather than all of them. Example 1

[0023] Figure 1 The figure is a flow chart of a charging station power control method combined with long-term monitoring and analysis of the present invention, wherein the method comprises:

[0024] Collect the charging equipment configuration and architecture of the target charging station and establish the station map structure.

[0025] Specifically, the charging equipment configuration and architecture of the target charging station are collected to establish a station map structure, which reflects the influence relationship between the charging equipment of the target charging station.

[0026] Specifically, to construct the station map result based on the graph neural network, first of all, it is necessary to collect data on the target charging station, including information such as the type, quantity, layout, and output power of the equipment. These data can be obtained in various ways, such as on-site measurements, equipment specification sheets, or from station operators. Then, the station graph structure is established with nodes representing devices and edges representing connections between devices. Among them, the weight of the edge can be set according to the output power of the device. For example, a charging pile with a larger output power has a greater impact on the station transformer or access end, and the corresponding weight is higher.

[0027] A timing monitoring and analysis channel is established, and the timing monitoring and analysis channel is used to evaluate the power conversion status of the charging station. The timing monitoring and analysis channel includes a long timing monitoring and analysis channel and a short timing monitoring and analysis channel.

[0028] Specifically, the timing monitoring and analysis channel is used to evaluate the power conversion state of the charging station from two time scales, and the timing monitoring and analysis channel includes a long timing monitoring and analysis channel and a short timing monitoring and analysis channel.

[0029] Among them, the long-term monitoring and analysis channel focuses on the power conversion characteristic fluctuations and power conversion states at a longer time scale, which can help understand the power conversion behavior and possible trends of the charging station over a longer time range (for example, one day, one week, or one month). It then reveals some periodic patterns, such as daily peaks and troughs, or specific patterns every week. Short-term analysis focuses on shorter time scales (for example, seconds, minutes, or hours). It is used to provide real-time or near real-time monitoring and analysis to promptly detect and respond to any real-time problems or changes.

[0030] In some embodiments, a timing monitoring and analysis channel is established, and the timing monitoring and analysis channel is used to evaluate the power conversion state of the charging station. The timing monitoring and analysis channel includes a long timing monitoring and analysis channel and a short timing monitoring and analysis channel, including:

[0031] Collect historical data of charging equipment and charging stations for cleaning preprocessing. Set the time accumulation rules and accumulation analysis features of long time series, and use historical data to train the time series analysis model to build the long time series monitoring and analysis channel. Set the sliding window size to configure the short cycle analysis rules and window analysis features, and use historical data to train the time series analysis model to build the short time series monitoring and analysis channel. Connect the long time series monitoring and analysis channel in parallel with the short time series monitoring and analysis channel to build the time series monitoring and analysis channel.

[0032] Specifically, first, extract historical operation data from the database of charging equipment and charging stations, such as charging power, equipment status, environmental conditions (temperature, humidity, etc.), and grid conditions. Exemplary data types include power curves, charging time, charging efficiency, load fluctuations, fault records, etc. Then, check the integrity of the data, fill in missing values, remove outliers, and smooth the noise data to ensure data quality and consistency.

[0033] Optionally, the acquired historical operation data is normalized, and variables of different dimensions are standardized (such as converting power, time, temperature and other data into a unified dimension) to facilitate subsequent model training and analysis.

[0034] Specifically, for long time series data (such as monthly, quarterly, and annual data), time accumulation rules and accumulation analysis features are defined, where the accumulation rule refers to the time window size of the accumulation statistics, and the accumulation analysis feature refers to the feature category to be accumulated. Optionally, the accumulation rules and accumulation analysis features correspond one to one, that is, different accumulation analysis features have different accumulation statistics window sizes.

[0035] Specifically, based on the time series model (such as ARIMA model, LSTM, GRU, etc.), combined with historical data, a long time series monitoring and analysis channel is constructed and trained. Exemplarily, first, feature engineering of historical data is performed according to the cumulative rules and cumulative analysis features, and appropriate features are selected or constructed. Exemplarily, for LSTM, the cumulative rules and cumulative analysis features are used as sliding window features. Then, the feature data obtained by feature engineering is used as training data to train the time series model and obtain a long time series monitoring and analysis channel. The long time series monitoring and analysis channel outputs trend prediction and status evaluation of the long-term power conversion state, which is used to identify the power conversion efficiency, fluctuations and potential efficiency decline trends of charging stations over a long period of time.

[0036] Specifically, the analysis period of short-time series data is set (such as minutes, hours, days, etc.), that is, the appropriate sliding window size (such as 5 minutes, 30 minutes, 1 hour, etc.) is selected according to the operation characteristics of the station. Then, the moving step of the sliding window is determined (such as every 5 minutes, every 10 minutes) to collect short-time series data in different time periods. Then, the short-term analysis rules and short-term analysis features are defined, wherein the short-term analysis rules define the allowable power conversion state range in the short period, exemplarily including real-time fluctuation rules (power fluctuations in a short period of time, used to determine whether there is an abnormal state beyond the normal fluctuation range), peak-valley change rules (peak and valley values ​​in the sliding window are set with reasonable upper and lower limits. When the peak-valley difference exceeds the preset threshold, the window is marked as unstable), power fluctuation rate rules (calculate the power change rate in the window, if the rate changes too fast, it is marked as a possible sudden state), etc.

[0037] Specifically, select a suitable time series analysis model (such as sliding window aggregation analysis, RNN, grey prediction model, etc.), and according to the same method principle as the above-mentioned construction and training of the long time series monitoring and analysis channel, input the short time series feature data for supervised training to obtain the short time series model. Then, apply the short time series model to the short time series data in the sliding window to generate a short time series monitoring channel for evaluating the power conversion state fluctuations of the charging station in the short term (such as hours or minutes) and the operating stability of the charging equipment.

[0038] Furthermore, the long-time series and short-time series analysis channels are connected in parallel to form a multi-scale time series monitoring and analysis channel. Through parallel connection, a comprehensive analysis of power conversion status at different time scales is achieved, so as to comprehensively monitor the operating status of the charging station from the long and short time dimensions.

[0039] Specifically, if the short-time channel detects frequent power fluctuations, it can be compared with the long-time channel to determine whether there is a long-term trend change. For example, if the short-time channel detects abnormal power fluctuations of a device, but the long-time channel shows a stable long-term trend, it may be a short-term interference, which needs to be judged in combination with the specific business scenario.

[0040] Through the above steps, the established time series monitoring and analysis channel can conduct a comprehensive and in-depth analysis of the power conversion status of the charging station in different time dimensions, and provide data support and decision-making basis for subsequent status prediction and optimization.

[0041] Furthermore, the cumulative analysis features include average power, peak power, fluctuation range, and equipment failure rate. The window analysis features include average power, maximum power, minimum power, and power fluctuation.

[0042] Specifically, the cumulative analysis features include at least average power, peak power, fluctuation range, and equipment failure rate. The window analysis features include at least average power, maximum power, minimum power, and power fluctuation.

[0043] Based on the site diagram structure, an equipment-independent analysis link and a site-global analysis link are established, and the equipment-independent analysis link and the site-global analysis link are respectively connected to the timing monitoring and analysis channel to obtain equipment-independent long-time / short-time power conversion evaluation and site-global long-time / short-time power conversion evaluation.

[0044] Specifically, the device independent analysis link focuses on the power conversion state and timing characteristics of a single charging device, and is used to analyze the independent operation of each device. The link is designed so that each device is an independent node in the graph structure, forming a separate device analysis link. This link only considers the power conversion, charging efficiency, fault conditions, etc. of the device itself, without considering the impact of the status of other devices on it.

[0045] Specifically, a long-term monitoring and analysis channel is used to analyze the historical data of a single device to evaluate the power conversion trend and health status of the device under long-term operation. Typical characteristic indicators include: charging power average, long-term fluctuations, long-term heating status, etc. Use a short-term monitoring and analysis channel to evaluate the power conversion fluctuations of the device within a short time range and identify abnormal conditions within a short period of time. Typical characteristic indicators include: instantaneous power fluctuation rate, short-term peak-to-valley changes, current shocks during charging, etc. Obtain independent long-term / short-term power conversion evaluation of the device.

[0046] Specifically, the global analysis link of the station analyzes the power conversion state of the entire station by considering the mutual influence between the devices and the overall operating efficiency through the correlation between the devices in the graph structure. Exemplarily, the link design includes constructing a global link in the graph structure (including each device node and its association), thereby reflecting the mutual influence relationship between each device in the station.

[0047] Specifically, the long-term changes in the overall power state of the station are analyzed through a long-term monitoring and analysis channel (such as a GNN-LSTM model based on LSTM and graph structure). For example, evaluate whether the long-term load changes of a transformer node affect the operating status of multiple charging piles connected to it. The generated global long-term evaluation results include: overall power conversion efficiency of the station, long-term power distribution uniformity, overall health score, etc. Use a short-term monitoring and analysis channel (such as a GNN-short-term model) to analyze power fluctuations and abnormal distributions in a short period of time to capture the overall operating status of the station in a short period of time. For example, when an abnormal fluctuation occurs in a charging pile, analyze whether it will cause instability in the power grid or power fluctuations of other equipment. Exemplarily, the generated global short-term evaluation results include: overall power fluctuation rate, short-term abnormal distribution, fault propagation path between devices, etc.

[0048] The above method steps, by introducing the station map structure, can establish a more comprehensive monitoring and analysis model, and then comprehensively analyze and evaluate the influence relationship between the equipment in the target station, which helps to more effectively manage and optimize the power of the charging station.

[0049] According to the independent long-time / short-time power conversion evaluation of the equipment and the global long-time / short-time power conversion evaluation of the station, the operating power of the charging equipment and the station is analyzed to identify abnormal risks.

[0050] In some embodiments, the operating power of the charging device and the station is analyzed based on the independent long-time / short-time power conversion evaluation of the device and the global long-time / short-time power conversion evaluation of the station to identify abnormal risks, including:

[0051] The time relationship between long and short time series is aligned by using the independent long-time / short-time power conversion evaluation of the equipment and the global long-time / short-time power conversion evaluation of the station, and the deviation of the long-time / short-time power conversion evaluation is calculated to obtain the deviation feature. The deviation features are arranged hierarchically according to the time alignment relationship, with the bottom layer being the deviation feature layer corresponding to the farthest time, and the top layer being the deviation feature layer corresponding to the most recent time. The features are aggregated upward in order from the bottom layer to the top layer, and the features of this layer are updated by using the aggregated features of the adjacent layers, and the final hierarchical aggregated features are obtained by continuous iteration. According to the final hierarchical aggregated features, an abnormal risk assessment is performed to obtain the abnormal risk.

[0052] Specifically, the equipment-independent long-time / short-time power conversion evaluation and the site-wide long-time / short-time power conversion evaluation results are extracted respectively. Then, according to the timestamp information of the sliding window corresponding to the evaluation, the time relationship between the long-time and short-time evaluation results is determined. For example, the long-time data (such as monthly data) and the short-time data (such as hourly data) are aligned according to the time nodes to find the same or adjacent time points. Then, for each time-aligned node, the difference between the equipment-independent long-time and short-time power conversion evaluations and the difference between the site-wide long-time and short-time power conversion evaluations are calculated to obtain the deviation characteristics of the corresponding time points. These deviation characteristics are the differences between long-time and short-time series data in different time periods, which are used to reveal the deviation changes between short-term fluctuations and long-term trends.

[0053] Specifically, the deviation features are arranged hierarchically according to the time dimension, preferably, from the farthest time point (such as a few months ago) to the closest time point (such as the current moment), wherein the deviation features at each time point are regarded as a deviation feature layer.

[0054] Furthermore, from the bottom layer to the top layer, feature aggregation and update are performed layer by layer. The aggregation rules can be weighted average, feature weighted combination or feature crossover. The updated features are used to update the status of this layer to ensure that the final hierarchical aggregation features can reflect the comprehensive status of long-term and short-term time series data within the overall time range.

[0055] Specifically, the final hierarchical aggregation feature is a comprehensive reflection of long-term and short-term time series data, which comprehensively evaluates the operating status of equipment or sites over the entire time range, including: power change trend, reflecting the long-term change pattern of power fluctuations and short-term fluctuation anomalies. Fluctuation amplitude and periodicity, including periodic fluctuation patterns, may be periodic faults or load fluctuations in equipment.

[0056] Furthermore, based on the deviation, fluctuation amplitude, periodicity and long-term trend in the hierarchical aggregation features, it is determined whether there is a potential abnormal risk. For example, if the long-term power trend of a device shown in the aggregation features continues to decline and deviates greatly from the short-term fluctuation, it may indicate that the device is at risk of performance degradation or power conversion efficiency decline. Preferably, the evaluation of abnormal risk is determined based on the abnormal trigger constraints in the expert system and the abnormal definition conditions of the target charging station.

[0057] Optionally, the risk level is graded. Normal corresponds to no significant deviation in the aggregated features and a stable power state. Slightly abnormal corresponds to small fluctuations in the aggregated features, which may be occasional fluctuations in a short period of time. Severe abnormal corresponds to significant deviations and fluctuations in the aggregated features, indicating long-term performance degradation or potential failures.

[0058] The above method steps calculate the deviation of long-time and short-time power conversion evaluation through time relationship alignment, and perform feature hierarchy arrangement and aggregation, which can effectively identify potential abnormal risks in the operating state.

[0059] In some implementations, the step of aggregating features upward in order from the lowest level to the highest level includes:

[0060] According to the time distance relationship of each layer, the weight of each layer is configured. Based on the weight of each layer, according to the deviation feature layer between the current layer and the subsequent adjacent time, feature aggregation is performed through a weighted average aggregation algorithm.

[0061] Optionally, a time weight coefficient is introduced during feature aggregation, so that the weight of the layer closer to the current moment is greater, thereby reflecting the time priority. Exemplarily, the weight is configured according to the inverse ratio of the time distance, so that the weight of the layer closer to the time distance is greater, and at the same time, the weights of all layers satisfy the weight sum of 1;

[0062] Exemplarily, the weight calculation is determined based on the following formula:

[0063] ;

[0064] in, is the weight. is the time distance between the i-th layer and the current time. The time distance is a relative distance, that is, the ratio of the time difference between the i-th layer and the current time to the farthest time difference.

[0065] Through the layer-by-layer weighted average aggregation algorithm, the features of different time levels can be effectively integrated, and the influence of recent features can be enhanced by introducing the time weight coefficient, so as to more comprehensively and accurately identify the power conversion state anomalies and potential risks of charging stations in long and short time series.

[0066] The control target is located according to the abnormal risk, and the control target is analyzed based on the abnormal risk to obtain a power control strategy, where the power control strategy is used to smooth or eliminate the abnormal risk.

[0067] Specifically, positioning and regulation analysis are carried out according to abnormal risks, so as to formulate appropriate power control strategies. Different regulation targets (independent equipment or charging stations) adopt corresponding analysis strategies and regulation methods to ensure the safe and stable operation of charging equipment and stations as a whole.

[0068] In some embodiments, locating a control target according to the abnormal risk, performing control analysis on the control target based on the abnormal risk, and obtaining a power control strategy includes:

[0069] When the control target is an independent device, the power deviation is determined according to the abnormal risk. With the power deviation as the adjustment target, the power control strategy is obtained based on the control parameters of the independent device. When the control target is a charging station, according to the adjustment relationship between the charging station and the independent device, with the power deviation as the adjustment target and the load balance of each device as the constraint condition, the power optimization adjustment of the independent device is performed to obtain the power control strategy.

[0070] Specifically, the control targets can be divided into two types: independent equipment and charging stations. Independent equipment refers to a single charging device or independent power module (such as a single charging pile, transformer or distribution box) in a charging station. For independent equipment, the main focus is on the power status, operating efficiency and risk of the equipment. If the control target is a charging station, it is necessary to focus on analyzing the load balancing and power distribution between devices based on the linkage relationship between the entire charging station and multiple devices inside it, and obtain a global power control strategy.

[0071] Specifically, for independent devices, that is, when the abnormal risk location points to a single device, a control analysis can be performed based on the power deviation of the device and its control parameters to formulate a power control strategy for the independent device. Each independent device has a set of control parameters (such as power adjustment range, response time, load capacity, control step, etc.), and the power deviation can be used to define the value range of the control parameters. For example, if the actual power of a charging pile is 110kW and the target power is 100kW, the power deviation is 10kW, and the control step can be set to 5kW, adjusting 5kW each time until the power drops to the target value of 100kW.

[0072] Specifically, when the control target is the entire charging station, it is necessary to consider the power adjustment relationship between each independent device and the load balancing problem of the station as a whole. For example, through the target (such as minimizing the sum of squares of power deviations) and combining iterative optimization with the optimization algorithm, the power allocation of each device is adjusted to determine the power control strategy.

[0073] Optionally, the overall power control strategy of the site may include: appropriately reducing the power of devices with too high power, appropriately increasing the power of devices with too low power, and determining the target power distribution ratio of each device. When some devices are overloaded, choose to shut down some devices or switch to backup devices. Adjust the load distribution relationship between devices to ensure the stability of power output.

[0074] By identifying and classifying the control targets, it is helpful to analyze and optimize the power control strategy according to their respective power deviations, improve the overall operating stability of the charging station, and reduce the impact of abnormal risks on station operations.

[0075] Furthermore, the method further comprises:

[0076] According to the station graph structure, the independent equipment is set as the sub-layer and the charging station is set as the parent layer, and the risk impact relationship between the sub-layer and the parent layer is established. According to the final hierarchical aggregation characteristics of the sub-layer and the final hierarchical aggregation characteristics of the parent layer, combined with the risk impact relationship, the abnormal risk identification of the charging station is performed.

[0077] Specifically, abnormal risks are identified through the final hierarchical aggregation characteristics of the sub-layer and the parent layer, as well as the risk impact relationship between the two. First, based on the station graph structure, all independent devices in the station are composed of sub-layers, and all independent devices are regarded as basic units in the sub-layer. At the same time, the entire charging station is regarded as a parent layer node, which is used to represent the overall operating status, power distribution and load management of the station. Then, through the risk impact relationship, multiple basic units in the sub-layer are connected to the parent layer.

[0078] For example, when an abnormality occurs in an independent device (such as a charging pile), its abnormal risk will be propagated to the parent layer (station) according to the performance load characteristics of the independent device, thereby affecting the power balance and load distribution of the entire station. The more devices the independent device is connected to, the greater the design load ratio of the independent device, and the greater the risk impact of the parent node. Preferably, the risk impact of each child node on the parent node is quantified by a risk impact coefficient, which can be calculated based on the power output, load ratio, and risk propagation probability of the independent device.

[0079] Specifically, based on the risk impact relationship, the final aggregation characteristics of the sub-layer and the final aggregation characteristics of the parent layer, abnormal risk identification of the charging station is performed, including calculating the risk deviation amount between the final aggregation characteristics of the sub-layer and the final aggregation characteristics of the parent layer. When the deviation amount exceeds a preset threshold, it indicates that certain risk states in the sub-layer have been transmitted to the parent layer, thereby causing abnormal risk of the station as a whole.

[0080] Furthermore, when abnormal risks are identified, corresponding early warning signals are generated to warn of abnormalities, and control strategies (such as load balancing adjustment, equipment start and stop, etc.) are formulated to eliminate or alleviate abnormal conditions.

[0081] The above method steps, through the hierarchical division of the station map structure and the establishment of risk impact relationships, can more accurately identify potential abnormal risks in charging stations, help formulate corresponding risk control strategies, and thus improve the overall operational safety and stability of charging stations.

[0082] In summary, the charging station power control method combined with long-term monitoring and analysis provided by the present invention has the following technical effects:

[0083] By collecting the charging equipment configuration and architecture information of the target charging station, a graph structure model of the station is established; a timing monitoring and analysis channel is constructed to monitor and evaluate the power conversion state of the charging station, wherein the timing monitoring and analysis channel includes a long-time monitoring channel and a short-time monitoring channel; based on the station graph structure, an independent equipment analysis link and a station global analysis link are established respectively, and these links are connected to the timing monitoring and analysis channel respectively, so as to obtain the independent long-time / short-time power conversion evaluation for the equipment and the global long-time / short-time power conversion evaluation for the station; based on the independent equipment and the global long-time / short-time power conversion evaluation for the station, the overall operating power of the charging equipment and the station is analyzed, and potential abnormal risks are identified; the target equipment or area is located according to the identified abnormal risks, and the regulation and analysis is performed based on the risk characteristics, and a power control strategy is formulated to smooth power fluctuations or eliminate potential abnormal risks. Thus, the technical effect of improving load adaptability and simplifying control strategies is achieved. Example 2

[0084] Figure 2 This is a schematic diagram of the structure of the charging station power control device combined with long-term monitoring and analysis of the present invention. For example, Figure 1 The flow chart of the charging station power control method combined with long-term monitoring and analysis in the present invention can be shown as follows: Figure 2 The structure shown is implemented.

[0085] Based on the same concept as the charging station power control method combined with long-term monitoring and analysis in the above embodiment, the present invention also provides a charging station power control device combined with long-term monitoring and analysis, including:

[0086] The graph structure building module 11 is used to collect the charging equipment configuration and architecture of the target charging station and establish the station graph structure.

[0087] The timing analysis building module 12 is used to establish a timing monitoring and analysis channel, which is used to evaluate the power conversion status of the charging station. The timing monitoring and analysis channel includes a long timing monitoring and analysis channel and a short timing monitoring and analysis channel.

[0088] The analysis link establishment module 13 is used to establish a device-independent analysis link and a site-global analysis link based on the site diagram structure, and connect the device-independent analysis link and the site-global analysis link to the timing monitoring and analysis channel respectively to obtain a device-independent long-time / short-time power conversion evaluation and a site-global long-time / short-time power conversion evaluation.

[0089] The abnormality identification module 14 is used to analyze the operating power of the charging equipment and the station according to the independent long-time / short-time power conversion evaluation of the equipment and the global long-time / short-time power conversion evaluation of the station, and identify abnormal risks.

[0090] The analysis and control module 15 is used to locate the control target according to the abnormal risk, perform control analysis on the control target based on the abnormal risk, and obtain a power control strategy, wherein the power control strategy is used to smooth or eliminate the abnormal risk.

[0091] The timing analysis building module 12 includes:

[0092] The data collection and preprocessing unit is used to collect historical data of charging equipment and charging stations for cleaning preprocessing.

[0093] The long time series analysis model training unit is used to set the time accumulation rules and accumulation analysis characteristics of the long time series, and use historical data to train the time series analysis model to build the long time series monitoring and analysis channel.

[0094] The short time series analysis model training unit is used to set the sliding window size to configure the short-cycle analysis rules and window analysis features, and use historical data to train the time series analysis model to build the short time series monitoring and analysis channel.

[0095] The time series monitoring and analysis channel construction unit is used to connect the long time series monitoring and analysis channel and the short time series monitoring and analysis channel in parallel to construct the time series monitoring and analysis channel.

[0096] Furthermore, in the timing analysis construction module 12, the cumulative analysis features include average power, peak power, fluctuation range, and device failure rate. The window analysis features include average power, maximum power, minimum power, and power fluctuation.

[0097] In some embodiments, the anomaly identification module 14 includes:

[0098] The time relationship alignment unit is used to use the independent long-time / short-time power conversion evaluation of the equipment and the global long-time / short-time power conversion evaluation of the station to align the time relationship of the long time and short time, calculate the deviation of the long-time / short-time power conversion evaluation, and obtain the deviation characteristics.

[0099] The deviation feature hierarchical arrangement unit is used to arrange the deviation feature hierarchies according to the time alignment relationship, the bottom layer is the deviation feature layer corresponding to the farthest time, and the top layer is the deviation feature layer corresponding to the most recent time.

[0100] The feature aggregation unit is used to aggregate features upward in the order from the bottom layer to the highest layer, update the features of this layer using the aggregated features of the adjacent layers, and continuously iterate to obtain the final hierarchical aggregated features.

[0101] The abnormal risk assessment unit is used to perform abnormal risk assessment according to the final hierarchical aggregation features to obtain the abnormal risk.

[0102] In some implementations, the feature aggregation unit in the anomaly identification module 14 includes:

[0103] The weight configuration unit is used to configure the weight of each layer according to the time distance relationship of each layer.

[0104] The feature weighted aggregation unit is used to perform feature aggregation based on the weights of each layer and according to the deviation feature layer between the current layer and the subsequent adjacent time through a weighted average aggregation algorithm.

[0105] In some embodiments, the analysis and control module 15 includes:

[0106] The independent device power control unit is used to determine the power deviation amount according to the abnormal risk when the control target is an independent device.

[0107] The power control strategy generating unit is used to obtain the power control strategy based on the control parameters of the independent device by taking the power deviation as the adjustment target.

[0108] The station-level power optimization and adjustment unit is used to optimize the power of independent equipment when the control target is a charging station, based on the adjustment relationship between the charging station and the independent equipment, with the power deviation as the adjustment target and the balancing of the load of each device as the constraint condition, to obtain the power control strategy.

[0109] Furthermore, the device also includes:

[0110] The risk impact relationship establishing unit is used to set the independent equipment as the sub-layer and the charging station as the parent layer according to the station diagram structure, and establish the risk impact relationship between the sub-layer and the parent layer.

[0111] An abnormal risk identification unit is used to identify abnormal risks of charging stations according to the final hierarchical aggregation characteristics of the sub-layer and the final hierarchical aggregation characteristics of the parent layer in combination with the risk impact relationship.

[0112] It should be understood that the embodiments mentioned in this specification focus on their differences from other embodiments. The specific embodiments in the aforementioned embodiment one are also applicable to the charging station power control device combined with long-term monitoring and analysis described in embodiment two. For the sake of brevity of the specification, they will not be further elaborated here.

[0113] It should be understood that the embodiments disclosed in the present invention and the above description can enable those skilled in the art to use the present invention to implement the present invention. At the same time, the present invention is not limited to the above-mentioned embodiments. It should be understood that those skilled in the art can still modify the technical solutions recorded in the above-mentioned embodiments, or replace some of the technical features therein by equivalents; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the embodiments of the present invention, and should be included in the protection scope of the present invention.

Claims

1. A charging station power control method combined with long-term monitoring and analysis is characterized in that: include: Collect the charging equipment configuration and architecture of the target charging station, and establish a station graph structure, wherein the station graph structure reflects the influence relationship between the charging equipment of the target charging station, wherein the station graph structure is established based on a graph neural network, with nodes representing equipment and edges representing connections between equipment; Establishing a time sequence monitoring and analysis channel, which is used to evaluate the power conversion status of the charging station, and includes a long time sequence monitoring and analysis channel and a short time sequence monitoring and analysis channel; Based on the station graph structure, establish a device-independent analysis link and a station-global analysis link, connect the device-independent analysis link and the station-global analysis link to the timing monitoring and analysis channel respectively, and obtain the device-independent long-time / short-time power conversion evaluation and the station-global long-time / short-time power conversion evaluation. The device-independent analysis link focuses on the power conversion state and timing characteristics of a single charging device, and is used to analyze the independent operation of each device; the station-global analysis link analyzes the power conversion state of the entire station through the correlation between devices in the graph structure, taking into account the mutual influence between devices and the overall operation efficiency; Analyze the operating power of the charging equipment and the station based on the independent long-time / short-time power conversion evaluation of the equipment and the global long-time / short-time power conversion evaluation of the station to identify abnormal risks; Locate a control target according to the abnormal risk, perform control analysis on the control target based on the abnormal risk, and obtain a power control strategy, wherein the power control strategy is used to smooth or eliminate the abnormal risk; The operation power of the charging equipment and the station is analyzed according to the independent long-time / short-time power conversion evaluation of the equipment and the global long-time / short-time power conversion evaluation of the station to identify abnormal risks, including: The time relationship between the long time series and the short time series is aligned by using the independent long time series / short time series power conversion evaluation of the equipment and the global long time series / short time series power conversion evaluation of the station respectively, and the deviation of the long time series / short time series power conversion evaluation is calculated to obtain the deviation feature, which is the difference between the long time series and the short time series data in different time periods, and is used to reveal the deviation change between the short-term fluctuation and the long-term trend; The deviation features are arranged hierarchically according to the time alignment relationship, with the bottom layer being the deviation feature layer corresponding to the farthest time, and the top layer being the deviation feature layer corresponding to the most recent time; Aggregate features upward in order from the bottom layer to the highest layer, use the aggregated features of the adjacent layers to update the features of this layer, and continuously iterate to obtain the final hierarchical aggregated features; Performing abnormal risk assessment according to the final hierarchical aggregation features to obtain the abnormal risk; Also includes: According to the station diagram structure, the independent equipment is set as the sub-layer and the charging station is set as the parent layer, and the risk impact relationship between the sub-layer and the parent layer is established; According to the final hierarchical aggregation features of the sub-layer and the final hierarchical aggregation features of the parent layer, combined with the risk impact relationship, abnormal risk identification of the charging station is performed; Wherein, locating a control target according to the abnormal risk, performing control analysis on the control target based on the abnormal risk, and obtaining a power control strategy includes: When the control target is an independent device, determining a power deviation amount according to the abnormal risk; Taking the power deviation as an adjustment target and based on the control parameters of the independent device, obtaining the power control strategy; When the control target is a charging station, the power of the independent device is optimized and adjusted according to the adjustment relationship between the charging station and the independent device, with the power deviation as the adjustment target and the load balancing of each device as the constraint condition, to obtain the power control strategy.

2. The charging station power control method combined with long-term monitoring and analysis as claimed in claim 1 is characterized in that: The upward feature aggregation in order from the lowest level to the highest level includes: According to the time distance relationship of each level, configure the weight of each level; Based on the weights of each layer, feature aggregation is performed using a weighted average aggregation algorithm according to the deviation feature layer between the current layer and the subsequent adjacent time.

3. The charging station power control method combined with long-term monitoring and analysis as claimed in claim 1 is characterized in that: The establishing of a timing monitoring and analysis channel, which is used to evaluate the power conversion status of the charging station, includes a long timing monitoring and analysis channel and a short timing monitoring and analysis channel, including: Collect historical data of charging equipment and charging stations for cleaning pre-processing; Setting the time accumulation rules and accumulation analysis features of the long time series, and using historical data to train the time series analysis model to build the long time series monitoring and analysis channel; The sliding window size is set to configure the short-cycle analysis rules and window analysis features, and the time series analysis model is trained using historical data to build the short-time series monitoring and analysis channel; The long time series monitoring and analysis channel and the short time series monitoring and analysis channel are connected in parallel to construct the time series monitoring and analysis channel.

4. The charging station power control method combined with long-term monitoring and analysis as claimed in claim 3 is characterized in that: The cumulative analysis features include average power, peak power, fluctuation range, and equipment failure rate; the window analysis features include average power, maximum power, minimum power, and power fluctuation.

5. A charging station power control device combined with long-term monitoring and analysis is characterized in that: The device is used to execute the charging station power control method combined with long-time monitoring and analysis according to any one of claims 1 to 4, and the device includes: A graph structure building module, which is used to collect the charging equipment configuration and architecture of the target charging station and establish a station graph structure, which reflects the influence relationship between the charging equipment of the target charging station. The station graph structure is established based on a graph neural network, with nodes representing devices and edges representing connections between devices; A timing analysis building module, which is used to establish a timing monitoring and analysis channel, which is used to evaluate the power conversion status of the charging station, and includes a long timing monitoring and analysis channel and a short timing monitoring and analysis channel; An analysis link establishment module, the analysis link establishment module is used to establish a device-independent analysis link and a station-global analysis link based on the station graph structure, and respectively connect the device-independent analysis link and the station-global analysis link to the timing monitoring and analysis channel to obtain a device-independent long-time / short-time power conversion evaluation and a station-global long-time / short-time power conversion evaluation. The device-independent analysis link focuses on the power conversion state and timing characteristics of a single charging device, and is used to analyze the independent operation of each device; the station-global analysis link analyzes the power conversion state of the entire station through the correlation between devices in the graph structure, taking into account the mutual influence between devices and the overall operation efficiency; An abnormality identification module, which is used to analyze the operating power of the charging equipment and the station according to the independent long-time / short-time power conversion evaluation of the equipment and the global long-time / short-time power conversion evaluation of the station, and identify abnormal risks; An analysis and control module is used to locate a control target according to the abnormal risk, perform control analysis on the control target based on the abnormal risk, and obtain a power control strategy, wherein the power control strategy is used to smooth or eliminate the abnormal risk.

Images