Edge computing-enabled peak-valley energy storage and charging system

Abstract

This invention relates to the field of intelligent charging technology, specifically to an edge computing-enabled peak-valley energy storage charging system. The system includes a grid frequency discrimination module, an energy storage scheduling output module, a load trend identification module, a multi-station matching and sorting module, and a load dynamic allocation module. Through continuous data analysis of grid frequency, energy storage unit power and remaining energy, and load current, this invention enables the system to accurately identify the linkage relationships of various parameters during frequency anomaly responses. It refines the condition determination for peak-valley energy storage triggering and scheduling, optimizes the energy storage output decision path, and enhances the real-time performance and flexibility of energy allocation among multiple stations through dynamic screening of load trends and changing stations. Furthermore, it improves the system's adaptability to complex supply and demand changes in station matching and priority sorting, dynamically allocates energy storage resources and loads, effectively improves power utilization and the collaborative control level of the energy storage system, and promotes the intelligent upgrading of energy management.

Description

Technical Field

[0001] This invention relates to the field of intelligent charging technology, and in particular to an edge computing-enabled peak-valley energy storage and charging system. Background Technology

[0002] The field of smart charging involves the efficient supply and dynamic distribution of electricity, including power system optimization encompassing electric vehicle charging, energy storage management, and energy dispatch. It covers the management and control of charging power, charging time periods, energy utilization efficiency, and charging / discharging sequence. Through the coordinated operation of power conversion and energy storage devices, it improves the utilization rate of electricity and the operational efficiency of charging infrastructure. Among these, traditional edge computing-enabled peak-valley energy storage charging systems refer to systems that, based on edge computing technology, collect and analyze data locally on key aspects such as load forecasting, energy dispatch, and energy storage control of charging stations. Combined with peak-valley electricity pricing strategies, these systems utilize energy storage devices to charge and store energy during periods of low grid load and discharge energy during peak periods to alleviate grid pressure and optimize energy distribution at charging stations.

[0003] Existing technologies rely on fixed parameters for scheduling in real-world scenarios, lacking a comprehensive assessment of the linkage between real-time grid anomalies and changes in the status of energy storage units. When encountering large-scale load fluctuations or changes in energy demand at multiple sites, the response strategy lags behind, failing to dynamically adjust energy storage triggering and output conditions based on parameters. Load trends and anomaly point identification are not timely, and energy allocation paths are rigid, resulting in insufficient energy storage response and load balancing capabilities. This limits the flexibility and efficiency of system operation, making it difficult to achieve adaptive allocation of diverse energy sources and real-time resource optimization when facing complex operating conditions. Summary of the Invention

[0004] The purpose of this invention is to address the shortcomings of existing technologies by proposing an edge computing-enabled peak-valley energy storage and charging system.

[0005] To achieve the above objectives, the present invention adopts the following technical solution: an edge computing-enabled peak-valley energy storage and charging system, the system comprising:

[0006] The power grid frequency discrimination module analyzes the changes in the output power, remaining power and load current of the energy storage unit based on the power grid frequency signals synchronously collected by the edge intelligent computing nodes of each site, compares the trends of frequency and energy storage parameters, judges the linkage characteristics and monitors the abnormal recovery process, and obtains the abnormal frequency linkage characteristics.

[0007] Based on the abnormal frequency linkage characteristics, the energy storage scheduling output module filters the energy storage unit output power, releaseable power and number, determines whether the releaseable power meets the peak and valley energy supply conditions, sorts the numbers and issues output instructions, and counts the power and current of the energy supply unit to obtain the scheduled energy storage output.

[0008] Based on the dispatched energy storage output, the load trend identification module compares the power load of each site with the load change trend over a period of time, filters sites with abnormal load change trends, sorts the load detection data of abnormal sites, analyzes the difference in load change of the top two sites in the sort, and obtains the load change range of the site.

[0009] The multi-site matching and sorting module calculates the remaining energy storage and electricity load of the site based on the load variation range of the site, compares the matching relationship, sorts them according to the priority of matching degree, and dynamically adjusts the sorting structure to obtain the priority site matching set.

[0010] The present invention is improved in that the abnormal frequency linkage feature includes disturbance type category, response peak amplitude, linkage duration period, the scheduled energy storage output includes energy storage output amplitude, scheduling period identifier, and scheduling unit set, the site load change range includes the maximum change amount of the range, change rate level, and associated site number, and the priority site matching set includes sorting weight parameter, site priority identifier, and matching sequence number.

[0011] The present invention is improved in that the power grid frequency discrimination module includes:

[0012] The frequency trend extraction submodule analyzes the time-series data of the output power, remaining power and load current of the energy storage unit synchronously collected by the edge intelligent computing nodes of each site based on the grid frequency signal collected synchronously by each site. By calculating the frequency change amplitude and direction at each time node, it compares the relationship between the frequency trend and the energy storage output at different stages to obtain the difference in the power frequency linkage trend.

[0013] The energy storage status linkage analysis submodule analyzes the changes in energy storage output power, remaining power and load current at the same site over multiple cycles based on the difference in the frequency power linkage trend. It calculates the directional consistency of each indicator and obtains the range of energy storage condition linkage by filtering the periodic data that meet the conditions.

[0014] The frequency anomaly evolution detection submodule analyzes the frequency data sequence based on the amplitude of the energy storage condition linkage interval, determines the continuous changes in energy storage output power and load current during the start and end points of the anomaly stage, calculates the response degree of energy storage power and load current to frequency disturbances, and obtains the abnormal frequency linkage characteristics by combining the disturbance type and periodic characteristics.

[0015] The present invention is improved in that the energy storage scheduling output module includes:

[0016] Based on the abnormal frequency linkage characteristics, the energy storage power screening submodule analyzes the output power, number and release capacity of the energy storage units synchronously collected by each edge computing node, compares the current frequency change trend with the number power response characteristics, determines which energy storage units have the conditions to respond to scheduling, and screens the numbers to participate in scheduling to obtain the number power linkage set.

[0017] The releaseable power determination submodule calculates the releaseable power corresponding to each energy storage unit based on the numbered power linkage set, analyzes the release capacity required during the current peak-valley operation phase, filters energy storage numbers that meet the energy supply conditions, optimizes the classification of energy storage units, and obtains a set of dispatchable energy storage numbers.

[0018] The scheduling instruction generation submodule compares the output power, release capacity and frequency response of each number according to the set of schedulable energy storage numbers, optimizes the priority order of the numbers, calculates the unified output amplitude of energy storage, adjusts the output mode of each number, issues scheduling instructions and monitors the response of each energy storage unit to obtain the scheduled energy storage output.

[0019] The present invention is improved in that the load trend identification module includes:

[0020] The load trend acquisition submodule analyzes the power load monitoring data of the site based on the dispatched energy storage output, compares the changes in power load curves over continuous periods, filters out key load trend segments within the period, and obtains the key load fluctuation curve.

[0021] The abnormal trend screening submodule judges the trend curve changes of each site based on the key load fluctuation curve, compares and finds sites with key anomalies in the rate of change, optimizes the trend curves of the top two sites, and obtains the site anomaly amplitude sequence.

[0022] The variation range calculation submodule compares the trend amplitude of the top two stations within the same period based on the abnormal amplitude sequence of the stations, analyzes the differences in load changes between the corresponding numbers of the two stations, determines the boundary of their variation range, and obtains the load variation range of the stations.

[0023] The present invention is improved in that the multi-station matching and sorting module includes:

[0024] The energy storage load comparison submodule compares the current remaining energy storage of each site with the electricity load of the same period based on the load variation range of the site, calculates the difference between the energy storage and the electricity load of the same period, and then performs synchronous normalization processing on the maximum load change range and the energy storage difference to obtain the energy storage load adaptation difference range.

[0025] The matching degree calculation submodule calculates the load change rate and energy storage release rate of the site during the same period based on the energy storage load adaptation difference, compares the load distribution differences between the sites, and obtains the matching degree level between the sites.

[0026] The priority sorting submodule analyzes the matching degree and site number of each screened site based on the matching degree level between the sites, and adjusts the priority of each site one by one according to the matching degree level to optimize the order of matching sites and obtain a priority site matching set.

[0027] The present invention has an improvement, wherein the system further includes:

[0028] Based on the priority site matching set, the load dynamic allocation module analyzes the energy storage output and load distribution of the target site, determines the priority site as the control target, adjusts the energy storage output mode, compares the load distribution changes before and after the adjustment, and obtains the load allocation change magnitude.

[0029] The load allocation change range includes the adjusted load value, the target site for allocation, and the allocation batch number.

[0030] The present invention is improved in that the dynamic load allocation module includes:

[0031] The priority site identification submodule analyzes the sorting order of each site and the current energy storage output and load distribution based on the priority site matching set, determines the site with the best priority and whose energy storage and load characteristics meet the control conditions, compares the priority and controllability among multiple sites, and obtains the control target site number.

[0032] The energy storage mode adjustment submodule analyzes the current energy storage output mode and energy storage response characteristics of the target site according to the target site number, optimizes the adjustment mode to adapt to the real-time load distribution, compares the changes in historical energy storage response with the current state, determines the optimal energy storage output mode, and obtains the energy storage output type.

[0033] The load amplitude calculation submodule analyzes the load distribution data before and after adjustment based on the energy storage output type, calculates the range of load change caused by the adjustment of the energy storage method, compares the load distribution shift brought about by the adjustment, filters and judges the degree of load redistribution, and obtains the load distribution change amplitude.

[0034] Compared with the prior art, the advantages and positive effects of the present invention are as follows:

[0035] In this invention, by continuously analyzing data on grid frequency, energy storage unit power and remaining energy, and load current, the system can accurately identify the linkage relationships of various parameters during frequency anomaly responses, refine the condition determination for peak-valley energy storage triggering and scheduling, optimize the energy storage output decision path, enhance the real-time performance and flexibility of energy allocation among multiple sites through dynamic screening of load trends and changing sites, further improve the system's adaptability to complex supply and demand changes in site matching and priority ranking, dynamically allocate energy storage resources and loads, effectively improve the efficiency of power utilization and the collaborative control level of the energy storage system, promote the intelligent upgrade of energy management, and facilitate adaptive energy balance under different operating scenarios. Attached Figure Description

[0036] Figure 1 This is a system flowchart of the present invention;

[0037] Figure 2 This is a flowchart of the power grid frequency discrimination module in this invention;

[0038] Figure 3 This is a flowchart of the energy storage scheduling output module in this invention;

[0039] Figure 4 This is a flowchart of the load trend recognition module in this invention;

[0040] Figure 5 This is a flowchart of the multi-station matching and sorting module in this invention;

[0041] Figure 6 This is a flowchart of the load dynamic allocation module in this invention. Detailed Implementation

[0042] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention.

[0043] In the description of this invention, it should be understood that the terms "length," "width," "upper," "lower," "front," "rear," "left," "right," "vertical," "horizontal," "top," "bottom," "inner," and "outer," etc., indicating orientation or positional relationships, are based on the orientation or positional relationships shown in the accompanying drawings and are only for the convenience of describing the invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation, and therefore should not be construed as a limitation of the invention. Furthermore, in the description of this invention, "a plurality of" means two or more, unless otherwise explicitly specified.

[0044] Example

[0045] Please see Figure 1 This invention provides a technical solution: an edge computing-enabled peak-valley energy storage and charging system comprising:

[0046] The power grid frequency discrimination module analyzes the changes in the output power, remaining power and load current of the energy storage unit based on the power grid frequency signals synchronously collected by the edge intelligent computing nodes of each site. It compares the changing trends between the current power grid frequency and the on-site energy storage parameters, determines whether the linkage characteristics between each parameter meet the peak-valley energy storage start-up standard, and monitors the continuous changes in the differences between each parameter during the frequency anomaly occurrence and recovery phases to obtain the abnormal frequency linkage characteristics.

[0047] Based on the abnormal frequency linkage characteristics, the energy storage scheduling output module filters the output power, releaseable power and number of the energy storage units associated with the edge intelligent computing nodes, determines whether the current releaseable power of the energy storage unit meets the peak and valley energy supply conditions, sorts the filtered energy storage unit numbers and issues output commands, and then counts the output power and current data of the energy supply units to obtain the scheduled energy storage output.

[0048] The load trend identification module compares the power load of each site with the load change trend over a period of time based on the dispatched energy storage output, filters sites with abnormal load change trends, sorts the load detection data of abnormal sites, analyzes the difference in load change of the top two sites in the sort, and obtains the load change range of the site.

[0049] The multi-site matching and sorting module calculates the remaining energy storage and electricity load data of related sites based on the load variation range of the sites, compares the matching relationship between the remaining energy storage and the load, prioritizes each site according to the degree of matching, and dynamically adjusts the sorting structure to obtain a priority site matching set.

[0050] The load dynamic allocation module analyzes the real-time energy storage output and load distribution of the target site based on the priority site matching set, determines the site with the highest priority as the control target, adjusts the energy storage output mode of the site, and compares the changes in the load distribution of the site before and after the adjustment to obtain the magnitude of the load allocation change.

[0051] The abnormal frequency linkage characteristics include disturbance type category, response peak amplitude, linkage duration period; the dispatched energy storage output includes energy storage output amplitude, dispatch period identifier, and dispatch unit set; the site load change range includes the maximum change amount in the range, change rate level, and associated site number; the priority site matching set includes sorting weight parameters, site priority identifier, and matching sequence number; and the load allocation change amplitude includes the adjusted load value, allocation target site, and allocation batch number.

[0052] In the power grid frequency discrimination module, edge computing nodes refer to edge computing terminal devices deployed at charging stations, energy storage stations, and other sites for local high-speed processing, analysis, and decision-making. They can collect, process, and issue control commands in real time. The power grid frequency signal refers to the actual operating frequency of the power grid detected by the power grid acquisition equipment in the power system, used to determine the current stable state of the power grid. Energy storage units refer to battery system modules or supercapacitors and other energy storage devices with independent control and monitoring capabilities; each unit can participate in energy storage management and scheduling independently. The trend of change refers to the direction and magnitude of change of a parameter (such as frequency, power, load, etc.) over a period of time, such as increase, decrease, or fluctuation. The linkage characteristic refers to the interaction of multiple parameters (…). The interaction between parameters such as frequency, energy storage power, and load current is a key indicator. For example, the synergistic phenomenon of increased output power of energy storage units when the grid frequency decreases is a key indicator. Peak-valley energy storage start-up criteria refer to the criteria and benchmarks used to determine whether energy storage units should start charging and discharging operations during peak or valley periods. These criteria are usually based on a combination of real-time parameters (such as frequency fluctuations and power margin). Frequency anomaly refers to the phenomenon where the grid operating frequency exceeds the safe operating range, indicating a supply-demand imbalance, impact, or fault. Differences between parameters during the recovery phase refer to the real-time changes and mutual response characteristics of various monitoring parameters (such as frequency, energy storage output, and load current) during the process of eliminating frequency anomalies and returning to the normal range.

[0053] In the energy storage dispatch output module, the releaseable power refers to the remaining electrical energy that the energy storage unit can currently use for output, to meet charging needs, or to support the power grid; the number refers to the unique identification code of the energy storage unit, used by the dispatch system to identify, sort, and operate specific units; the peak and valley power supply conditions refer to the rules and standards set according to the system strategy to determine whether the energy storage unit can (or needs) to output or charge during peak hours (high load) or valley hours (low load); the output command refers to the control command issued by the system to the energy storage unit, requiring it to perform charging or discharging operations at a set power within a specific time.

[0054] In the load trend identification module, the power load of each site refers to the total amount of electricity currently consumed by each charging station, energy storage station, or local microgrid, reflecting the current power demand status of the site; the load change trend refers to the growth, decline, or fluctuation trajectory of the power load within a specified time period, used to reflect the dynamic demand of the site; abnormal sites refer to sites whose power load change trends deviate from the normal range, exhibit sudden changes, or drastic fluctuations, as discovered through analysis; load detection data refers to real-time and historical data information on power load collected by sensors and acquisition devices; and the load change difference refers to the comparison results of the power load change amplitude between different sites or different time periods, used to identify sites with large change amplitudes.

[0055] In the multi-site matching and sorting module, associated sites refer to all charging stations, energy storage stations, or microgrid nodes in the current system that need to participate in energy storage scheduling and load balancing; the matching relationship between remaining energy storage and load refers to the mutual adaptation between the remaining dispatchable energy storage of each site and the current power load, reflecting the site's ability to bear new loads or release energy storage; the matching degree refers to the adaptability of the remaining energy storage to the load demand obtained through calculation, which is usually measured by standards such as ratio and correlation; the dynamic adjustment of the sorting structure refers to the operation process of automatically reordering the priority and resource allocation order of each site according to the real-time changes in the system's operating status.

[0056] In the load dynamic allocation module, real-time energy storage output refers to the charging and discharging power and continuous output status of energy storage units at each site at the current moment; load distribution refers to the current distribution of electrical load among multiple sites, reflecting the overall system load balance level; control target refers to the specific site selected by the system as the target for this round of scheduling and output adjustment; energy storage output mode refers to the working status of the energy storage unit, including specific operation modes such as charging, discharging, and standby; load distribution change refers to the change in the load distribution pattern and values ​​among sites before and after system adjustment, used to analyze the scheduling effect and load balance level.

[0057] Please see Figure 2 The power grid frequency discrimination module includes:

[0058] The frequency trend extraction submodule analyzes the time-series data of the output power, remaining power and load current of the energy storage unit synchronously collected by the edge intelligent computing nodes of each site based on the grid frequency signal collected synchronously by each site. By calculating the frequency change amplitude and direction at each time node, it compares the relationship between the frequency trend and the energy storage output at different stages to obtain the difference in the power frequency linkage trend.

[0059] The frequency values, energy storage unit output power values, remaining power values, and load current values ​​collected from each site are used to form corresponding time-series data sequences at a sampling frequency of one time point per second. Each parameter sequence needs to be precisely aligned according to the timestamp to ensure data consistency and parallelism. In actual sampling, for example, with an analysis cycle of 300 seconds, each data set is 300 records long. Then, in the frequency sequence of each site, the absolute value of the frequency difference between two adjacent time points is taken as the frequency change amplitude. If the current value is greater than the previous time point value, the direction of that period is upward; otherwise, it is downward. If the absolute value of the difference is lower than the set fluctuation identification benchmark value, such as 0.05Hz, the segment is marked as a stable trend. Next, under the stable frequency trend division, the change amplitude of energy storage output power, remaining power, and load current for the corresponding time period is extracted, represented by the difference between the current value and the previous value, forming a time-segment parameter change sequence consistent with the frequency trend. The change in energy storage output power and frequency in each segment is then analyzed. The correlation is determined by comparing the direction of power change with the direction of frequency change within each segment. When both change in the same direction, it is considered positive linkage; when they change in opposite directions, it is considered negative linkage; and when there is no significant change, it is considered weak linkage. The degree of linkage is assessed by counting the number of time points in each segment where the direction of frequency change is consistent with the direction of other parameter changes. For example, if 240 out of 300 time points show consistent changes in the three parameters, the consistency ratio is 80%, which can be classified as strong linkage; if only 120 time points show consistent changes, the consistency ratio is 40%, which is classified as weak linkage. Five levels of classification standards can be set according to this method: more than 80% is extremely strong linkage, 65%-80% is strong linkage, 50%-65% is moderate linkage, 35%-50% is weak linkage, and less than 35% is extremely weak linkage. Finally, the linkage situation of each station in different time periods is classified according to the level, and the differences in the electrical frequency power linkage trend of different stations in different frequency trend stages are output.

[0060] The energy storage status linkage analysis submodule analyzes the changes in energy storage output power, remaining power and load current at the same site over multiple cycles based on the difference in the linkage trend of electric frequency power, calculates the directional consistency of each indicator, and obtains the range of energy storage condition linkage by filtering the periodic data that meet the conditions.

[0061] Within each linked trend cycle, the output power, remaining power, and load current of the energy storage units at the current site are extracted for that time period, forming a three-column data list with time as the horizontal axis. Then, for each second of the three parameters, the direction of increase or decrease at adjacent time points is calculated. By judging the relationship between the current value and the previous value, if all three increase or decrease simultaneously, they are considered to be in the same direction; otherwise, they are considered to be incompatible. The number of time points with consistent direction throughout the entire cycle is counted, and this value is divided by the total cycle duration to obtain the direction consistency ratio. When this ratio exceeds 65%, it indicates that the changes of the three parameters in that cycle show a synchronous trend; if it is below 65%, the cycle is discarded without further processing. Subsequently, the difference between the maximum and minimum values ​​of each parameter in the direction-consistent cycle is calculated, representing the fluctuation amplitude of that parameter. The fluctuation amplitude of the three parameters is multiplied by a preset weighting factor, with energy storage power having the highest weight, followed by remaining power, and then load current, forming a weighted sum. This value reflects the strength of the linkage amplitude within a cycle, with a threshold of 1.5. If the weighted sum is greater than this value, the cycle is considered a strong linkage cycle of energy storage status; if it is lower than this value, the linkage amplitude is considered insignificant and is not included in the analysis. For example, in a certain site's cycle, the maximum power change is 1.5 kW, the power change is 3%, and the load current change is 1.2 A. After substituting the weights, the weighted linkage amplitude is calculated to be 1.89, which is higher than the set threshold. Therefore, this cycle is retained as a valid analysis segment. The time periods in multiple cycles where all sites meet the conditions of directional consistency and linkage amplitude are integrated, and the linkage interval amplitude of the energy storage conditions corresponding to the time period is output.

[0062] The frequency anomaly evolution detection submodule analyzes the frequency data sequence based on the amplitude of the energy storage condition linkage interval, and determines the continuous changes in energy storage output power and load current during the start and end points of the anomaly phase using the following formula:

[0063]

[0064] The response of energy storage power and load current to frequency disturbances is calculated. Combined with the disturbance type and periodic characteristics, the abnormal frequency linkage characteristic GS is obtained, where Q... i Q represents the energy storage output power at time i. i-1 J represents the energy storage output power at time i-1. i J represents the load current at time i. i-1 R represents the load current at time i-1. j R represents the grid frequency at time j. j-1 Let denot be the power grid frequency at time j-1, δ be the smallest positive number to prevent the denominator from being zero, λ be the frequency disturbance fluctuation coefficient, and n be the number of sampling periods during the abnormal phase.

[0065] The abnormal frequency linkage characteristic refers to the comprehensive response strength of the energy storage unit output power and load current relative to the change in grid frequency during periods of abnormal grid frequency (such as large fluctuations or exceeding the stable range). The larger the characteristic value, the more drastic the response changes of energy storage and load during periods of abnormal frequency, and the closer the correlation. The smaller the value, the less sensitive or less affected the energy storage and load are.

[0066] Q i J represents the energy storage output power at time i (unit: kilowatts). i R represents the load current at time i (in amperes). j Let represent the grid frequency at time j (in Hertz), δ be a tiny positive number to prevent the denominator from being zero (set to 0.0001 here), λ be the frequency disturbance fluctuation coefficient (set to 2.4 in the experiment), and n be the number of sampling periods during the abnormal phase (currently set to 4). The sampling data are as follows: energy storage output power is 280, 290, 305, and 300 kW; load current is 420, 430, 440, and 435 kW; grid frequency is 49.8, 49.6, 49.4, and 49.7 kW. After normalization, the processed energy storage power sequence is:

[0067] Q = [0.56, 0.58, 0.61, 0.60];

[0068] The load current sequence is as follows:

[0069] J = [0.84, 0.86, 0.88, 0.87];

[0070] The distance between each set of sampling points in the molecule is calculated as follows:

[0071] Group 1:

[0072]

[0073] Group 2:

[0074]

[0075] Group 3:

[0076]

[0077] The sum of the numerators is:

[0078] ∑=0.0283+0.0361+0.0141=0.0785;

[0079] Calculate the denominator; the squared difference between each frequency sampling segment is:

[0080] First paragraph: (49.6-49.8) 2 =0.04;

[0081] Second paragraph: (49.4-49.6) 2 =0.04;

[0082] Third paragraph: (49.7-49.4) 2 =0.09;

[0083] Adding δ = 0.0001, the denominator becomes:

[0084] ∑=0.04+0.04+0.09+0.0001=0.1701;

[0085] Substituting into the formula, we get:

[0086]

[0087] The results show that during the frequency anomaly phase, there is a quantitative correlation between the synchronous variation amplitude of energy storage output power and load current in the time series and the intensity of frequency disturbance. The value GS = 1.1078 indicates that the joint response capability of the energy storage system to the grid frequency anomaly disturbance is at a relatively strong level, reflecting that the site has a high potential for energy storage regulation participation and load support during the current anomaly period. This value can be further used as a core reference quantity to judge the priority response sites in the subsequent dispatch output and load trend identification phase. It can also be used in the multi-site matching and sorting module to establish a response intensity sorting mechanism and clarify the priority control targets of energy storage under the anomaly disturbance situation.

[0088] Please see Figure 3 The energy storage dispatch output module includes:

[0089] The energy storage power screening submodule analyzes the output power, number, and releaseable power of energy storage units synchronously collected by each edge computing node based on the abnormal frequency linkage characteristics. It compares the current frequency change trend with the number power response characteristics to determine which energy storage units have the conditions to respond to scheduling, and screens the numbers that participate in scheduling to obtain the number power linkage set.

[0090] The edge computing nodes at each site collect real-time data on the current output power, unique ID, and releaseable electricity of each energy storage unit. After summarizing the data for each site's energy storage units, the grid frequency change trend for the corresponding time period is extracted and compared with the power response characteristics exhibited by that unit during historical frequency disturbances. For example, if a certain energy storage unit's output power rapidly increases to over 90% of its maximum output power within 5 seconds when the frequency drops, it is determined to have a fast response capability. If the power change rate is less than 30% of the maximum output power per second, it is considered to have a weak response capability. A threshold is set to classify units with high response capability, and the direction of the current frequency change is further analyzed. If the frequency is decreasing, energy storage numbers with outstanding power response in historical frequency decrease scenarios are selected and included in the dispatchable candidate set. Conversely, if the frequency is increasing, numbers that can actively reduce power output in the context of increasing frequency are matched. The process of determining whether each energy storage unit meets the current dispatch conditions also requires comparing the output power response time with the duration of the frequency change trend. For example, if the frequency decrease duration is 10 seconds, and the response time of a certain energy storage number needs to be greater than 15 seconds, then this number will be excluded and will not enter the dispatch response judgment. Then, all energy storage numbers that pass the above judgment conditions are filtered and output to form a set of numbers with power linkage that meet the frequency trend response conditions.

[0091] The releaseable power determination submodule calculates the releaseable power corresponding to each energy storage unit based on the numbered power linkage set, analyzes the release capacity required during the current peak-valley operation phase, filters energy storage numbers that meet the energy supply conditions, optimizes the classification of energy storage units, and obtains a set of dispatchable energy storage numbers.

[0092] The system reads the current releaseable power value of each energy storage unit in the set one by one. This value is obtained by subtracting the minimum protection power from the product of the current state of charge and the total capacity of the energy storage unit. For example, if a battery has a total capacity of 100kWh, a current state of charge of 85%, and a minimum protection power of 10kWh, then the releaseable power is 75kWh. Next, the scheduling system determines whether the current operating period is a peak or off-peak period. If it is a peak period, power needs to be output to support the high load. If it is an off-peak period, the scheduling target changes to energy storage charging. The minimum energy supply capacity required for scheduling during peak periods is set to 300kWh. If the energy storage units in the current linkage set... If the total releaseable energy is less than 300 kWh, some energy storage units with a releaseable energy of less than 20 kWh need to be removed. Then, the set is replenished by re-searching for energy storage units that are not in the set but have medium response capability and a releaseable energy of more than 30 kWh. In addition, during the judgment process, it is also necessary to determine whether there is a risk of over-discharge based on the rate of change of energy in each energy storage unit over the past 30 minutes. If the rate of energy decrease exceeds twice its average rate, it will not be included in the dispatchable objects. After re-screening the energy storage units in the set through this process, the energy storage units that meet the energy supply threshold requirements are regrouped and output according to the numbers to obtain the set of dispatchable energy storage units that can participate in the dispatch at the current stage.

[0093] The scheduling instruction generation submodule compares the output power, releaseable energy, and frequency response of each scheduling energy storage number based on the set of available numbers, and optimizes the number priority order using the following formula:

[0094]

[0095] Calculate the unified output amplitude SR of energy storage, adjust the output mode of each numbered unit, issue dispatch commands and monitor the response of each energy storage unit to obtain the dispatched energy storage output, where PR z ER represents the output power of the z-th energy storage unit. z Rf represents the releasable electrical energy of the z-th energy storage unit. z Rf represents the grid frequency currently monitored by the z-th energy storage unit. s The reference frequency, n, represents the current stage. SR This represents the total number of energy storage units included in this dispatch.

[0096] The unified output amplitude of energy storage is the total energy release intensity that the overall energy storage system can currently output when multiple energy storage units cooperate in peak-valley scheduling. It is achieved by comprehensively considering the current output power, releaseable electricity, and real-time response characteristics of each energy storage unit to changes in grid frequency. All energy storage units with scheduling capabilities are weighted and aggregated according to priority. It reflects the energy release level that all currently schedulable energy storage units can uniformly output after cooperating, based on their respective states and system requirements. It is a key comprehensive indicator for system scheduling decisions and issuing output commands.

[0097] To compare the current response status of the output power, releaseable energy, and grid frequency corresponding to each number, all parameters need to be processed according to a unified dimension to calculate the unified output amplitude SR of energy storage and the output power parameter PR. z This indicates the instantaneous power level available for each number, and the amount of electricity that can be released (ER). z This indicates the remaining discharge capacity of the numbered device, and the power grid monitoring frequency Rf. z The reference frequency is Rf, which is the frequency data collected by the current device. s The frequency is uniformly set to 50Hz, based on the target of normalized and stable operation of the power system. To maintain the comparability of various physical quantities on a uniform scale, normalization is required. Here, the normalized values ​​are directly used in the calculation, as detailed below:

[0098] For energy storage device A1, PR z =120, ER z =400, Rf z =49.8, after normalization are 0.800, 0.842, and 0.496 respectively; for A3, PR z =95, ER z =300, Rf z =49.5, after normalization are 0.633, 0.707, and 0.492 respectively; for A4, PR z =130, ER z =450, Rf z =50.0, after normalization, they are 0.867, 0.894, and 0.500 respectively. Substituting them into the formula:

[0099] A1:

[0100] A3:

[0101] A4:

[0102] The calculation result is:

[0103] SR≈0.731+0.528+0.820=2.079;

[0104] This value represents the relative strength of the uniform output amplitude of the currently schedulable energy storage devices after calculation. Under the normalized benchmark, if the scheduling threshold is set to 2, the current result 2.079 is higher than the threshold, indicating that the scheduling conditions are met. The system will adjust the output status of each numbered energy storage device according to priority and issue corresponding power commands item by item as the input basis for subsequent load trend identification and dynamic allocation. The formula introduces a dynamic suppression mechanism for device synchronization by using frequency deviation as the denominator adjustment term. It combines the product of power and energy to construct a multi-dimensional index of output potential, avoiding the scheduling deviation caused by a single weight mechanism, thereby improving the overall control accuracy and interpretability.

[0105] Please see Figure 4 The load trend recognition module includes:

[0106] The load trend acquisition submodule analyzes the power load monitoring data of the site based on the dispatch energy storage output, compares the changes in power load curves over continuous periods, and filters out key load trend segments within the period to obtain the key load fluctuation curve.

[0107] The system reads the energy storage output power of each site and the electricity load monitoring data collected by the load sensor within the time period. The load data for each site is summarized into a time-series sequence, one record per minute, forming an hourly load change array of length 60. This is then divided into 12 analysis periods, each 5 minutes long. Within each period, the difference between the initial and final load values ​​is calculated and compared with the difference between the maximum and minimum load values ​​within the period to determine the severity of the load change. A benchmark value for judging fluctuations is set at 10% of the maximum load value. If the difference between the maximum and minimum values ​​within a period exceeds this benchmark value, and the net change rate from start to end exceeds 5%, then a fluctuation is judged to be significant. To identify critical load fluctuations, complete load trend curves are extracted from the identified fluctuation segments. The rising, falling, and drastic fluctuation segments of the current load are then organized according to time sequence. Combined with the records of changes in the energy storage output power, it is compared whether the increase in energy storage power occurs synchronously in the load rising segment within the same time period. If the energy storage output increases from 20kW to 40kW within 5 minutes while the load increases from 200kW to 240kW, it is recorded as a consistent trend segment. If there are two or more consistent trend segments, the hourly load trend curve of that site is marked as a critical curve. By integrating the direction, amplitude, and continuity of the critical segments, the critical load fluctuation curves of each site are output.

[0108] The abnormal trend screening submodule is based on the key load fluctuation curve to judge the change of the trend curve of each site, compares and finds sites with key anomalies in the rate of change, optimizes the trend curves of the top two sites, and obtains the site anomaly amplitude sequence.

[0109] For each station, the curve is compared segment by segment, and the rate of change value is extracted. That is, the load increase or decrease per minute is used as the basis for the calculation of the rate of change. The maximum rate of change within each 5-minute segment in 60 minutes is extracted and used as the representative rate of that period segment. Then, the maximum rate values ​​in 12 segments of each station are sorted, and the segment with the maximum rate is selected as the abnormal point segment. It is determined whether the rate exceeds the set abnormal rate benchmark value of 20kW / min. If it exceeds, it is marked as a critical abnormal point. If more than three segments of the rate are found to be greater than the threshold in a certain station, the entire station is regarded as an abnormal station. The maximum rates of all abnormal stations are re-sorted from high to low, and the top two stations are selected as the key trend stations. The trend curves corresponding to these two stations are refined, and the combination structure of each inflection point, steep rise and fall segment and gradual change area in the curve is extracted. The rate of change of all key segments is re-summarized, and an integrated rate sequence is output. The abnormal amplitude sequence of the station is constructed in chronological order.

[0110] The variation range calculation submodule compares the trend amplitude of the top two stations within the same period based on the abnormal amplitude sequence of the stations, analyzes the difference in load changes between the corresponding numbers of the two stations, determines the boundary of their variation range, and obtains the load variation range of the stations.

[0111] Read the abnormal rate sequences of the top two stations and pair them time-by-time. Compare the difference in rate of change within the same period, extract and mark time points where the difference is greater than 10 kW / min, and then extract the corresponding load change values ​​at those time points. Extract the start and end values ​​of that time period from the load curves of the two stations respectively, and compare their changes to further determine whether the change is within the normal fluctuation range. Set the threshold for the normal load fluctuation range to ±30 kW. If the difference in load change between the two stations exceeds this value, it is recorded as a strong fluctuation segment in that period. Continue to mark the strong fluctuation segment according to its start and end times, and define the... The segment is the critical range of load change. By checking the difference between the basic load value and the upper limit load of the two stations corresponding to their numbers, it is determined whether there is a capacity redundancy difference. For example, the basic load of station A01 is 150kW, the upper limit is 200kW, and the current maximum value of the fluctuation segment is 195kW. The basic load of station B03 is 140kW, the upper limit is 220kW, and the current maximum value is 180kW. The difference is significant. Based on the combined judgment of the difference in the rate of change and the difference in the load change amplitude in each period, the maximum value, minimum value and duration of the two stations in the critical segment are extracted, and the complete station load change range is output.

[0112] Please see Figure 5 The multi-station matching and sorting module includes:

[0113] The energy storage load comparison submodule compares the current remaining energy storage of each site with the electricity load during the same period based on the load variation range of the site, calculates the difference between the energy storage and the electricity load during the same period, and then performs synchronous normalization processing on the maximum load variation range and the energy storage difference to obtain the energy storage load adaptation difference range.

[0114] Extract the current remaining energy storage value and the average electricity load for each station within the specified interval. Pair the remaining energy storage value with the load value station-by-station and perform a difference calculation. The difference is the remaining energy storage minus the average load. A positive difference indicates that the station has load-bearing capacity; a negative difference indicates that the current energy storage cannot cover the corresponding load demand. For example, station A has 180kWh of remaining energy storage and a concurrent load of 150kWh, resulting in a difference of 30kWh; station B has 120kWh of remaining energy storage and a concurrent load of 170kWh, resulting in a difference of -50kWh. Then, extract the maximum load change amplitude for each station within the current fluctuation interval, defined as the maximum increase or decrease in load per unit time for that station. Normalize this maximum change value with the aforementioned energy storage and load difference. The normalization process is performed by uniformly calculating the maximum load change amplitude across all stations. The scale is reduced to the 0 to 1 range, and the corresponding energy storage difference is processed in the same way. During the normalization process, the minimum benchmark value is set as the smallest negative load change value among all sites, and the maximum benchmark value is the maximum positive change value. After normalization, the absolute value of the difference between the two normalized indicators is taken and then compared and evaluated. For example, the normalized value of the load change amplitude of site A is 0.75, and the normalized value of the energy storage difference is 0.85, with a difference of 0.10. The corresponding values ​​for site B are 0.95 and 0.40, with a difference of 0.55. Therefore, the adaptation difference of site B is significantly higher. According to the preset adaptation difference range classification standard, the difference in the range of 0.0-0.2 is defined as high adaptation, 0.2-0.4 is medium adaptation, and more than 0.4 is severe adaptation deviation. According to this classification standard, the adaptation differences of all sites are classified and sorted, and the corresponding energy storage load adaptation difference amplitude of each site is output.

[0115] The matching degree calculation submodule calculates the load change rate and energy storage release rate of each site during the same period based on the magnitude of the difference in energy storage load adaptation, and compares the load distribution differences between sites using the following formula:

[0116]

[0117] Obtain the matching degree between sites, where MD o DE indicates the relevance level of site o. o DL represents the remaining energy stored at station o.o DR represents the current electrical load of station o. o DV represents the energy release rate of site o within a specified time period. o D represents the rate of load variation at site o within a specified time period. ov This represents the difference in load distribution between site o and site v, n MD This represents the total number of sites participating in the matching analysis;

[0118] The matching degree level between sites refers to an index used to measure the adaptability between a site's energy storage and load under the combined effects of multiple parameters such as current remaining energy storage, power load, energy storage release rate, load change rate, and load distribution differences with other sites. The higher the value, the worse the adaptability between the site's energy storage and load, the greater the supply and demand difference between sites, and the relatively lower the priority.

[0119] The load variation rate and energy storage release rate of a site during the same period are calculated. These two rates are extracted as the power change amplitude within a time interval divided by the duration of the period, forming the dynamic response magnitude within the cycle. Furthermore, based on the load distribution of the site's region, the load mean difference between the site and other sites within the same scheduling cycle is calculated, and the sum of the squares of this difference represents the site's distributional disparity in global scheduling. After dimension conversion for each parameter, the normalized result is directly incorporated into the formula calculation. Taking site A as an example, its remaining energy storage DE... o The current electrical load is DL, which is 220. o The value is 140, and the result after calculating the squared difference is (DE). o -DL o ) 2 =(220-140) 2 =6400, the energy storage release rate DR of site A in the current cycle o The load variation rate (DV) is 18. o The value is 6. After taking the square root of the sum of squares, we get:

[0120]

[0121] The load distribution of this site differs from that of several other sites in the same period by 12, 10, and 14. Squaring these differences and summing them gives:

[0122]

[0123] Substitute into the formula:

[0124]

[0125] The results indicate that, within the current cycle, site A possesses sufficient energy storage resources, a relatively moderate electricity load, and high energy storage release efficiency and load regulation response capabilities. Furthermore, given minimal differences in load distribution compared to other sites, it exhibits good overall supply-demand matching stability. The calculated matching degree level MD... o The value of ≈14.56 is relatively low among all participating sites, indicating that it has strong comprehensive allocation capabilities and is suitable for priority inclusion in the target selection range for multi-site scheduling.

[0126] The priority sorting submodule analyzes the matching degree and site number of each screened site based on the matching degree level between sites, and adjusts the priority of each site one by one according to the matching degree level to optimize the order of matching sites and obtain the priority site matching set.

[0127] Obtain the adaptation difference range of all sites from the previous step, and associate each site number with its matching degree. Determine whether the matching degree level of each site falls into three ranges: high, medium, and low. The matching degree level classification standard is set as follows: adaptation difference range less than 0.2 is high level, 0.2 to 0.4 is medium level, and greater than 0.4 is low level. Then, filter all site numbers in the high level and initially sort them according to the remaining energy storage capacity from high to low. If there are multiple sites in the high level with the same remaining energy storage capacity or a difference of less than 5kWh, further sort them according to... The current load stability is judged, and the load stability is taken as the average load change rate not exceeding 5% over three consecutive cycles as the stability reference value. If the change rate of a certain site is 3% and that of another site is 6%, the former is ranked higher. The same sorting rules are applied to the medium-level sites, and only the high-level sorting results are arranged before the medium-level results. For low-level sites, only the number is kept for reference and not included in the matching queue. After all sorting is completed, a site priority list is formed. Each site in the list records the number, remaining energy storage, adaptation difference, and sorting position. The matching set of the priority sites is output.

[0128] Please see Figure 6 The load dynamic allocation module includes:

[0129] The priority site identification submodule analyzes the sorting order of each site and the current energy storage output and load distribution based on the priority site matching set, determines the site with the best priority and whose energy storage and load characteristics meet the control conditions, compares the priority and controllability among multiple sites, and obtains the control target site number.

[0130] The system reads the ranking position, remaining energy storage capacity, actual output power, and current load value for each site within the matching set. It determines whether the site has been in the top two for three consecutive rounds and exhibits high load stability, serving as one of the initial screening criteria. Stability is judged based on load changes not exceeding a set threshold of 5% within three adjacent cycles. Simultaneously, sites with remaining energy storage capacity exceeding their load are selected, prioritizing sites with a difference greater than 20 kWh for the next round of evaluation. The system then calculates the deviation between the current energy storage output and the current load for each site, using the ratio of actual energy storage output power to load as a control ratio reference. If this ratio is between 0.8 and 1.2, the controllability is considered manageable; otherwise, it is marked as weak. A baseline value of 0.2 is set for the deviation. The site with the smallest deviation among all sites is identified as having the best controllability. Through multiple rounds of comparison of ranking position, controllability indicators, and load synchronization capabilities, the site with the highest priority ranking and the smallest control deviation is identified as the target for this round of control. The corresponding site number is extracted, and the target site number is output.

[0131] The energy storage mode adjustment submodule analyzes the current energy storage output mode and energy storage response characteristics of the target site according to the target site number, optimizes the adjustment mode to adapt to the real-time load distribution, compares the changes in historical energy storage response with the current state, determines the optimal energy storage output mode, and obtains the energy storage output type.

[0132] Based on the target site number, the current operating mode of the energy storage unit at that site is extracted, including three states: charging, discharging, and maintaining output. Combined with the corresponding power output change records for that site, it is determined whether the direction of power change is consistent with the direction of load change within a 10-minute period. If there are more than three reverse changes or a response delay exceeding 2 minutes, the current energy storage output mode is considered incompatible with the load change. Statistical analysis is performed on the energy storage response characteristics of this site, including the correlation between energy storage output power adjustment speed, output duration, and frequency interference. A response delay threshold of 2 minutes is set; if the adjustment delay exceeds this value and the output fluctuates frequently, then... Recommendations for switching energy storage output modes include adjusting continuous output to intermittent segmented output, or setting a minimum output lock-in range during frequency disturbances. Extract energy storage operation records for the same time period of the previous day from historical comparisons, and analyze whether the energy storage control mode used under similar load conditions was more stable. If the historical energy storage power output deviation is lower than the current scheme, prioritize reverting to the historical control mode. If the historical control fluctuation amplitude is greater than the current state, maintain the current scheme and force a strategy switch only after the output fluctuation duration exceeds 15 minutes. Finally, determine the output mode that best matches the current load distribution and output the energy storage output type that the site should currently use.

[0133] The load amplitude calculation submodule analyzes the load distribution data before and after adjustment based on the energy storage output type, calculates the range of load change caused by the adjustment of the energy storage method, compares the load distribution shift brought about by the adjustment, filters and judges the degree of load redistribution, and obtains the load distribution change amplitude.

[0134] Read the power load records of each site within 5 minutes before and 5 minutes after the energy storage adjustment, compare the total power load of the site before and after the adjustment, and the load increase / decrease values ​​for each minute. Mark whether there is a phenomenon of load transfer to other sites during the process. By reading the synchronous fluctuation data of the load of multiple sites, determine whether there is a cross-site load compensation phenomenon caused by the adjustment of this site. If the load change direction of other sites is opposite to that of the adjusted site and the change exceeds 5kW, it is considered that the load has been redistributed. Extract the amplitude value of the redistribution point, count the load increase and decrease of all relevant sites within the adjustment window, and calculate the total load change range by accumulating the absolute values. Then compare the total change value with the original load change value of the adjusted site. Set the redistribution ratio threshold to 30%. If more than 30% of the total change is borne by other sites, the adjustment is considered a high-amplitude load distribution operation. If the ratio is between 15% and 30%, it is considered a medium amplitude. If it is below 15%, it is considered a slight distribution behavior. Output the load distribution change amplitude caused by this round of energy storage output adjustment.

[0135] The above are merely preferred embodiments of the present invention and are not intended to limit the present invention in any other way. Any person skilled in the art may make changes or modifications to the above-disclosed technical content to create equivalent embodiments that can be applied to other fields. However, any simple modifications, equivalent changes, and modifications made to the above embodiments based on the technical essence of the present invention without departing from the scope of the present invention shall still fall within the protection scope of the present invention.

Claims

1. Edge intelligence empowerment type peak-valley energy storage charging system, characterized in that, The system comprises: The grid frequency discrimination module analyzes the energy storage unit output power, the remaining power and the load current change based on the grid frequency signals synchronously collected by the edge intelligent calculation nodes of each site, compares the trends of the frequency and the energy storage parameters, judges whether the linkage characteristics between parameters meet the peak-valley energy storage starting standard, and monitors the abnormal recovery process to obtain abnormal frequency linkage characteristics; The energy storage scheduling output module filters the energy storage unit output power, the releasable power and the number based on the abnormal frequency linkage characteristics, judges whether the releasable power meets the peak-valley energy supply condition, sorts the numbers and issues the output instruction, and counts the energy supply unit power and the current to obtain the scheduling energy storage output; The load trend identification module compares the power consumption load of each site and the load change trend in a period of time based on the scheduling energy storage output, filters the sites with abnormal load change trend, sorts the abnormal site load detection data, analyzes the difference of the load change of the top two sites, and obtains the site load variation interval; The multi-site matching sorting module calculates the site remaining energy storage and power consumption load based on the site load variation interval, compares the matching relationship, sorts the matching degree priority, dynamically adjusts the sorting structure, and obtains the priority site matching set.

2. The edge intelligence enabled peak valley energy storage charging system of claim 1, wherein, The abnormal frequency linkage characteristics include the disturbance type category, the response peak amplitude and the linkage duration cycle, the scheduling energy storage output includes the energy storage output amplitude, the scheduling time period identifier and the scheduling unit set, the site load variation interval includes the interval maximum variation, the variation rate level and the associated site serial number, and the priority site matching set includes the sorting weight parameter, the site priority identifier and the matching sequence number.

3. The edge intelligence enabled peak valley energy storage charging system of claim 1, wherein, The grid frequency discrimination module comprises: The frequency trend extraction submodule analyzes the time sequence data of the energy storage unit output power, the remaining power and the load current synchronously collected by each site based on the grid frequency signals synchronously collected by the edge intelligent calculation nodes of each site, compares the relationship between the frequency trend of the difference stage and the energy storage output by calculating the frequency change amplitude and direction of each time node, and obtains the power linkage trend difference between the frequency and the power; The energy storage state linkage analysis submodule analyzes the changes of the energy storage output power, the remaining power and the load current in multiple cycles based on the power linkage trend difference between the frequency and the power, calculates the direction consistency of each index, and obtains the energy storage condition linkage interval amplitude by filtering the cycle data meeting the conditions; The frequency abnormal evolution detection submodule analyzes the frequency data sequence based on the energy storage condition linkage interval amplitude, judges the continuous change of the energy storage output power and the load current during the start and end points of the abnormal stage, calculates the response degree of the energy storage output power and the load current to the frequency disturbance, combines the disturbance type and the cycle characteristics, and obtains the abnormal frequency linkage characteristics.

4. The edge intelligence enabled peak valley energy storage charging system of claim 1, wherein, The energy storage scheduling output module comprises: The energy storage power filtering submodule analyzes the energy storage unit output power, the number and the releasable power synchronously collected by each edge intelligent calculation node based on the abnormal frequency linkage characteristics, compares the current frequency change trend and the number power response characteristics, judges which energy storage units have the response scheduling conditions, filters the numbers participating in the scheduling, and obtains the number power linkage set; The releasable power determination submodule calculates the releasable power of each energy storage unit according to the numbered power linkage set, analyzes the required release capacity in the current peak-valley operation stage, filters the energy storage numbers that meet the energy supply conditions, optimizes the classification of energy storage units, and obtains a set of dispatchable energy storage numbers; The dispatch instruction generation submodule compares the output power, releasable power and frequency response of each number according to the set of dispatchable energy storage numbers, optimizes the number priority order, calculates the unified output amplitude of energy storage, adjusts the output mode of each number, issues a dispatch instruction and monitors the response of each energy storage unit, and obtains the dispatchable energy storage output.

5. The edge intelligence enabled peak valley energy storage charging system of claim 1, wherein, The load trend identification module comprises: The load trend collection submodule analyzes the power consumption load monitoring data of the station based on the dispatchable energy storage output, compares the changes of the power consumption load curve in consecutive time periods, filters the load trend segments that fluctuate significantly within the period, and obtains the key load fluctuation curve; The abnormal trend filtering submodule judges the trend curve change of each station based on the key load fluctuation curve, compares and finds the stations with significant change rate, optimizes the trend curves of the top two stations, and obtains the station abnormal amplitude sequence; The variation interval calculation submodule compares the trend amplitudes of the top two stations within the same period according to the station abnormal amplitude sequence, analyzes the differences between the corresponding numbers of the two stations in load variation, judges the difference variation interval limits, and obtains the station load variation interval.

6. The edge intelligence enabled peak valley energy storage charging system of claim 1, wherein, The multi-station matching and sorting module comprises: The energy storage load comparison submodule compares the current residual energy storage and simultaneous power consumption load of each station based on the station load variation interval, calculates the difference between the synchronous energy storage and power consumption load, and then synchronously normalizes the maximum load change amplitude and the energy storage difference, to obtain the energy storage load adaptation difference amplitude; The matching degree calculation submodule calculates the load variation rate and energy storage release speed of the stations in the same period based on the energy storage load adaptation difference amplitude, compares the load distribution differences between the stations, and obtains the matching degree level between the stations; The priority sorting submodule analyzes the matching degree and station number of each filtered station based on the matching degree level between the stations, adjusts the priority of each station one by one according to the matching degree level, optimizes the arrangement order of the matching stations, and obtains the priority station matching set.

7. The edge intelligence enabled peak valley energy storage charging system of claim 1, wherein, The system further comprises: The load dynamic allocation module analyzes the energy storage output and load distribution of the target station based on the priority station matching set, judges the station with the optimal priority as the control object, adjusts the energy storage output mode, compares the load distribution changes before and after the adjustment, and obtains the load allocation change amplitude. The load allocation change amplitude comprises the load value after adjustment, the allocation target station, and the allocation batch number.

8. The edge intelligence enabled peak valley energy storage charging system of claim 7, wherein, The load dynamic allocation module comprises: The priority station identification submodule analyzes the sorting order of each station and the current energy storage output and load distribution based on the priority station matching set, judges the station with the optimal priority and whose energy storage and load characteristics meet the control conditions, and compares the priority and controllability of multiple stations to obtain the control target station number. The energy storage mode adjustment submodule analyzes the current energy storage output mode and energy storage response characteristics of the regulation target site according to the regulation target site serial number, optimizes the adjustment mode to adapt to the real-time load distribution, compares the historical energy storage response and the change of the current state, judges the optimal energy storage output mode, and obtains the energy storage output type; The load amplitude calculation submodule analyzes the load distribution data before and after adjustment according to the energy storage output type, calculates the load change range caused by the energy storage mode adjustment, compares the load distribution transfer caused by adjustment, screens and judges the degree of load redistribution, and obtains the load distribution change amplitude.

Images