Multi-functional cooperative control method and system for distributed energy storage charging

Abstract

The application discloses a kind of distributed energy storage charging multifunctional collaborative control method and system, it is related to electrical engineering technical field.The method includes: collecting the original data of grid side, energy storage side, charging side and photovoltaic side, the original data is carried out outlier rejection, protocol conversion and data compression processing, and output standardized data;Based on the standardized data, first carry out energy storage safety hard constraint check to determine safety state, then according to grid load rate and user charging urgency carry out scene threshold self-adaptation to output scene type and basic charging power, then generate preliminary power instruction through lightweight reinforcement learning algorithm, finally, the preliminary power instruction is judged, and the final power instruction is output.The application balances the demand of grid, charging, energy storage by constructing closed-loop control architecture of data acquisition, dynamic decision, power execution, feedback correction, and achieves the effect of improving grid interaction reliability and user charging experience.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of electrical engineering, in particular to a multifunctional collaborative control method and system for distributed energy storage charging. BACKGROUND

[0002] With the rapid development of distributed energy storage and new energy charging technology, the photovoltaic storage and charging integrated system has become an important part of the new power system and is widely used in industrial and commercial photovoltaic storage and charging integration, virtual power plants, vehicle-to-grid (V2G) and other scenarios. The existing photovoltaic storage and charging system adopts a fixed priority strategy. When the power grid issues peak shaving / frequency modulation instructions, the user charging is often interrupted forcibly. If the charging is prioritized, the power grid instruction response is insufficient. The traditional reinforcement learning algorithm uses redundant state parameters, resulting in data transmission delay and decision lag. The global random sampling of the experience replay pool is insufficient for learning in emergency scenarios. The existing scheme embeds SOC checks in the decision execution link, which may lead to overcharging of the energy storage, and does not combine SOH for hierarchical protection, shortening the life cycle of the distributed energy storage. When the sum of the grid instruction power and the charging demand power exceeds the available power of the energy storage, the existing technology adopts a "one-size-fits-all" strategy without clear quantitative reduction logic. SUMMARY

[0003] The present application aims to provide a multifunctional collaborative control method and system for distributed energy storage charging to solve the problems raised in the background.

[0004] To achieve the above-mentioned purpose, the present application provides the following technical scheme: a multifunctional collaborative control method for distributed energy storage charging, comprising the following steps:

[0005] Step S100: Collecting raw data from the power grid side, the energy storage side, the charging side and the photovoltaic side, performing outlier rejection, protocol conversion and data compression processing on the raw data, and outputting standardized data;

[0006] Step S200: Based on the standardized data, first perform a safety hard constraint check of the energy storage to determine the safety state, then perform scene threshold self-adaptation according to the power grid load rate and user charging urgency to output the scene type and basic charging power, then generate a preliminary power instruction through a lightweight reinforcement learning algorithm, and finally perform compliance judgment on the preliminary power instruction to output a final power instruction;

[0007] Step S300: Calculate the total power demand according to the final power instruction, distribute the energy storage power and adjust the charging power according to the scene type, and output the actual running data after execution;

[0008] Step S400, calculate the deviation value based on the actual operation data and the final power instruction, if the deviation value exceeds the preset threshold, adjust the energy storage power distribution parameter and push the state information to the power grid, the user and the next round of decision-making link, form a closed loop control.

[0009] Preferably, the step S100 further comprises the following sub-steps:

[0010] Step S110, collect the frequency deviation of the power grid side through the power grid data acquisition component , load rate and instruction power ;

[0011] Collect the state of charge , health state and rated power of each energy storage unit on the energy storage side through the energy storage data acquisition component, and then calculate the average state of charge , minimum health state and total available power , wherein and , i=1,2,…,N, N is the total number of energy storage units;

[0012] Collect the charging demand , Level1 user number and emergency user number of each charging pile on the charging side through the charging data acquisition component, and calculate the total charging demand , wherein , j=1,2,…,M, M is the total number of charging piles;

[0013] Collect the real-time output of photovoltaic side through the photovoltaic data acquisition component ;

[0014] Step S120, adopt 3σ criterion to eliminate data exceeding reasonable range in the original data, the data exceeding reasonable range including , or , or , , the eliminated data is replaced by the previous period data;

[0015] Convert the data in different protocol formats in the original data into IEC 61850 standard format; adopt differential compression processing for , , and , and the compression ratio is not less than 1:5;

[0016] Step S130, integrate the data processed in step S120, output the standardized data containing 、 、 、 、 、 、 、 、 、 .

[0017] Preferably, the step S110 further comprises the following sub-steps:

[0018] The step S111, through the frequency sensor, collects the real-time frequency of the power distribution network , the frequency deviation is calculated , 50Hz is the rated frequency of the power grid;

[0019] Through the voltage and current sensor, the real-time power of the 10kV bus is collected , combined with the bus rated capacity , the load rate is calculated ;

[0020] Through the grid dispatching interface, the instruction power is obtained , the instruction power , discharge is positive, charge is negative;

[0021] Step S112, through the CAN bus, collect the of N distributed energy storage units , calculate the , the ; Collect the of each energy storage unit , calculate the total available power ;

[0022] Step S113, through the charging protocol interface, collect the of M charging piles , calculate the total charging demand ; Through the user identification component, identify the user priority, and statistically obtain the number of Level1 users ;

[0023] Based on the user-reported estimated driving time , the number of emergency users is statistically obtained , the number of emergency users corresponds to Level1 users;

[0024] Step S114, collecting the real-time output of the photovoltaic inverter through a photovoltaic communication interface , the photovoltaic communication interface supports Modbus TCP protocol, and the data collection delay is not more than 30 ms.

[0025] Preferably, the step S200 further comprises the following sub-steps:

[0026] Step S210, if the , switch to the energy storage protection priority mode, prohibit all energy storage from participating in grid assistance, and only allocate basic charging power to Level 1 users , and the total charging power , 0.2 is the maximum allowed discharge depth;

[0027] If , allow the energy storage to participate in grid assistance, but the discharge power , 0.3 is the maximum allowed discharge depth, allocate basic charging power to Level 1 users , and suspend Level 2 user charging;

[0028] If , continue to execute step S220;

[0029] Step S220, adjusting the emergency threshold and normal threshold of the frequency deviation according to the load rate , outputting a grid emergency scenario or a grid normal scenario;

[0030] Adjusting the basic charging power of Level 1 users according to the emergency user proportion ;

[0031] Step S230, constructing a state space based on the 6-dimensional core parameters in the standardized data, the 6-dimensional core parameters including , , , , , ;

[0032] Calculating the reward value through a three-section reward function , wherein is the grid compliance reward, the grid responsiveness , , , , , , is the charging guarantee reward, wherein the Level 1 interruption rate , , when , when , is the energy storage protection penalty, wherein when , when , when ;

[0033] The algorithm model is trained by adopting a scenario hierarchical experience replay mechanism, and preliminary power instructions are output, the preliminary power instructions including instruction power in an emergency scenario of a power grid and preliminary charging power , and preliminary power grid auxiliary power in a normal scenario of the power grid ;

[0034] In step S240, it is checked whether the preliminary power instructions meet energy storage safety constraints and power grid standard constraints, the energy storage safety constraints being instruction total power , and the power grid standard constraints being power grid instruction power , is the maximum allowable response power of the power grid;

[0035] If yes, final power instructions are output;

[0036] If no, step S230 is returned to readjust the reward function weight and generate new preliminary power instructions, and the number of retries is less than or equal to 3 times. If the number of retries is still 3 times after 3 times of retries, the energy storage protection priority mode is triggered.

[0037] Preferably, the step S220 further includes the following sub-steps:

[0038] In step S221, if , the emergency threshold of is set to ±0.3 Hz, the normal threshold is set to ±0.2 Hz, and the response trigger delay is less than or equal to 20 ms;

[0039] If , the emergency threshold of is set to ±0.5 Hz, the normal threshold is set to ±0.3 Hz, and the response trigger delay is less than or equal to 30 ms;

[0040] If , the emergency threshold of is set to ±0.8 Hz, the normal threshold is set to ±0.5 Hz, and the response trigger delay is less than or equal to 50 ms;

[0041] In step S222, if , the basic charging power , the real-time output is preferentially allocated ;

[0042] If , the basic charging power , adopt and energy supply;

[0043] If , the basic charging power , the remaining energy capacity participates in grid assistance.

[0044] Preferably, the step S300 further comprises the following sub-steps:

[0045] Step S310, if it is a grid emergency scenario, the total discharge power demand , is the final grid instruction power, is the final charging power;

[0046] If , the power distribution coefficient is allocated to each energy unit power , wherein i=1,2,…,N;

[0047] If , the charging power reduction ratio is calculated, and the actual charging power is obtained;

[0048] If it is a normal grid scenario, the total power consumption demand , is the final grid assistance power;

[0049] If , the power is preferentially met, and the remaining power is taken as the actual grid assistance power ; if , the grid assistance power reduction ratio is calculated, and the actual grid assistance power is obtained;

[0050] Step S320, the energy storage control component issues the each energy unit power or to each energy unit, drives the energy storage inverter to adjust the power according to the PWM control signal, and outputs the actual grid response power ; the charging control component issues to each charging pile, adjusts the charging pile output current, and outputs the actual charging power ;

[0051] If an overcurrent or overtemperature anomaly is detected, the emergency stop switch is triggered, power distribution is suspended, and the system switches to emergency mode. The emergency mode only guarantees 5kW / pile charging for Level 1 users. The overcurrent is defined as current > 1.2 × rated current, and the overtemperature is defined as temperature > 65℃.

[0052] Preferably, step S400 further includes the following sub-steps:

[0053] Step S410: Calculate the power grid response deviation. ,like but ;

[0054] Calculate charging power deviation ,like but ;

[0055] Calculate photovoltaic output deviation The To contribute to the actual development of photovoltaics;

[0056] Step S420, if Then improve energy storage unit The coefficient is 10%, that is And return to step S200 for feedback. Correction factor;

[0057] like and Then from An additional 5kW / pile will be allocated to Level 1 users, and feedback will be returned to step S300. Fluctuation warning and reservation As backup power;

[0058] Upload to the power grid dispatch center With remaining total available power Push charging progress to users ,cost And abnormal alerts, push equipment health reports to the operation and maintenance terminal, the equipment health reports contain Decline trend Inverter temperature profile During the charging progress This is the amount of charge already applied. To meet user power needs, It is a time-of-use electricity price.

[0059] The present invention also provides a multi-functional collaborative control system for distributed energy storage charging, comprising:

[0060] A data acquisition module is configured to acquire raw data of a power grid side, an energy storage side, a charging side and a photovoltaic side, perform outlier rejection, protocol conversion and data compression processing on the raw data, and output standardized data.

[0061] A central decision module is configured to, based on the standardized data, first perform energy storage safety hard constraint checking to determine a safety state, then perform scene threshold self-adaptation according to a power grid load rate and a user charging urgency to output a scene type and a basic charging power, subsequently generate a preliminary power instruction through a lightweight reinforcement learning algorithm, and finally perform compliance judgment on the preliminary power instruction to output a final power instruction.

[0062] A distributed execution module is configured to calculate a total power demand according to the final power instruction, perform energy storage power distribution and charging power adjustment according to the scene type, and output actual operation data after execution.

[0063] A feedback module is configured to calculate a deviation value based on the actual operation data and the final power instruction, and if the deviation value exceeds a preset threshold, adjust an energy storage power distribution parameter and push state information to a power grid, a user and a next round of decision-making link to form a closed-loop control.

[0064] The application further provides an electronic device, which is a physical device, and the electronic device comprises:

[0065] a processor and a memory, wherein the memory is communicatively connected to the processor.

[0066] The memory is configured to store executable instructions executed by the at least one processor, and the processor is configured to execute the executable instructions to implement the multifunctional collaborative control method for distributed energy storage charging.

[0067] The application further provides a computer readable storage medium, wherein the computer readable storage medium stores a computer program, and the computer program is executed by a processor to implement the multifunctional collaborative control method for distributed energy storage charging.

[0068] Compared with the prior art, the application has the following beneficial effects:

[0069] By constructing a closed-loop control architecture of data acquisition, dynamic decision-making, power execution and feedback correction, the demands of a power grid, charging and energy storage are balanced, the reliability of power grid interaction and the charging experience of a user are improved, the decision-making efficiency and adaptability are optimized through a lightweight reinforcement learning algorithm and a scenario threshold self-adaptation, the system response delay is reduced, the multi-device compatibility is enhanced, the energy storage is prevented from over-discharging and health degradation through energy storage safety hard constraint checking and a hierarchical protection mechanism, the life cycle of energy storage is prolonged, and the equipment replacement cost is reduced. BRIEF DESCRIPTION OF DRAWINGS

[0070] Figure 1 A main flowchart of a distributed energy storage charging multifunctional cooperative control method provided for an embodiment of the present application is shown in the figure.

[0071] Figure 2 A structural schematic diagram of a distributed energy storage charging multifunctional cooperative control system provided for an embodiment of the present application is shown in the figure.

[0072] Figure 3 A structural schematic diagram of an electronic device provided for an embodiment of the present application is shown in the figure. DETAILED DESCRIPTION

[0073] The technical solutions in the embodiments of the present application will be described clearly and completely below with reference to the accompanying drawings in the embodiments of the present application. Obviously, the described embodiments are only some of the embodiments of the present application, but not all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative work fall within the protection scope of the present application.

[0074] Please refer to Figure 1 The present application provides a distributed energy storage charging multifunctional cooperative control method, which comprises the following steps.

[0075] In step S100, the original data of the power grid side, the energy storage side, the charging side and the photovoltaic side are collected, and the original data are subjected to outlier rejection, protocol conversion and data compression processing to output standardized data.

[0076] The original data of the power grid side refers to the initial data collected from the distribution network related to power operation, including but not limited to the real-time frequency, voltage, current of the distribution network, the peak shaving and frequency modulation instruction power issued by the power grid, etc. The original data of the energy storage side refers to the initial data collected from the distributed energy storage unit, including but not limited to the state of charge (SOC), the state of health (SOH), the rated power, the real-time operating temperature of each energy storage unit, etc. The original data of the charging side refers to the initial data collected from the charging pile and the user interaction link, including but not limited to the current charging demand power of each charging pile, the user priority identifier, the user's expected driving time, etc. The original data of the photovoltaic side refers to the initial data collected from the photovoltaic inverter, mainly the real-time output power of the photovoltaic system, the operating state of the photovoltaic module, etc.

[0077] Additionally, it should be noted that "outlier removal" refers to removing invalid data from the original data that exceeds the normal range, such as frequency deviations exceeding ±2Hz, or energy storage unit state of charge less than 0% or greater than 100%, which are clearly inconsistent with actual operating patterns, by using preset data rationality judgment rules to avoid abnormal data interfering with subsequent decision-making; "protocol conversion" refers to uniformly converting communication protocol data output from different acquisition objects and in different formats into a standard protocol format (such as the IEC 61850 standard format) that meets the parsing requirements of the system's central decision-making module, solving the protocol compatibility problem between different devices; "data compression processing" refers to reducing the amount of data through specific algorithms (such as differential compression algorithms), especially for slowly changing data (such as energy storage unit health status and rated power), reducing the pressure of data transmission and storage while ensuring data accuracy; "standardized data" refers to the final data set that has a unified format, accurate content, and no redundancy after outlier removal, protocol conversion, and data compression processing, and this data set can be directly called by the central decision-making module.

[0078] In one possible implementation, the data acquisition is performed by the data acquisition module in the system. This module connects to the power grid, energy storage, charging, and photovoltaic equipment through different acquisition components. For example, it acquires grid-side data through an IEC 61850 smart gateway, energy storage-side data through a CAN bus, charging-side data through a communication module supporting the OCPP 1.6 protocol, and photovoltaic-side data through a Modbus TCP protocol interface, ensuring real-time acquisition of data from all dimensions.

[0079] Step S100 further includes the following sub-steps:

[0080] Step S110: Collect the frequency deviation on the power grid side using the power grid data acquisition component. Load factor and command power ;

[0081] The state of charge of each energy storage unit on the energy storage side is collected by the energy storage data acquisition component. Health status and rated power Then the average state of charge is calculated. Minimum health status and total available power ,in, and In this context, i = 1, 2, ..., N, where N is the total number of energy storage units;

[0082] The charging data acquisition component collects the charging demand of each charging pile on the charging side. Level 1 user count and emergency user number , the total charging demand is calculated wherein, j=1, 2, …, M, M is the total number of charging piles;

[0083] The real-time output of the photovoltaic side is collected through a photovoltaic data collection component ;

[0084] In step S120, the 3σ criterion is used to eliminate data outside the reasonable range in the original data, including , or , or , is the rated power of photovoltaic, and the data after elimination is replaced by the previous period data;

[0085] The data in different protocol formats in the original data are converted into IEC 61850 standard format; the , , and are processed by differential compression, and the compression ratio is not less than 1:5;

[0086] In step S130, the data processed in step S120 are integrated, and the standardized data containing , , , , , , , , , are output.

[0087] Among them, the “3σ criterion” refers to eliminating abnormal data (such as frequency deviation >±2Hz, energy storage <0% or >100%) outside the range of “mean±3 times standard deviation” based on the normal distribution characteristics of data, to ensure data validity; the “IEC61850 standard format” refers to the general data exchange format in the field of power system automation, which is used to unify the protocols of different devices such as power grid, energy storage, and charging, to avoid data parsing errors caused by incompatible protocols; the “differential compression processing” refers to storing only the difference between the current value and the previous period value for , and other slowly changing data (update period 1min), to realize data compression ratio not less than 1:5 and reduce transmission pressure.

[0088] In a possible implementation, in addition to adopting the 3σ criterion, the outlier rejection can also adopt the "adjacent cycle comparison method" for sudden extreme data (such as photovoltaic output dropping to 0) - if the difference between the current data and the average data of the previous 3 cycles exceeds 50%, it is determined as an abnormality and the previous cycle data is replaced to further improve the data reliability; in addition to supporting the IEC 61850 format, the protocol conversion link can also be compatible with the Modbus RTU format, adapting to the communication needs of some old charging piles, without the need for additional deployment of protocol converters.

[0089] For example, in a possible implementation, the power grid data acquisition component acquires the real-time frequency of the power distribution network through a frequency sensor = 50.35 Hz (calculated = +0.35 Hz), acquires the real-time power of the 10kV bus through a voltage and current sensor = 9.2 MW (the rated capacity of the bus , calculated = 9.2 MW), and obtains the peak shaving instruction through a power grid dispatching interface = 200 kW; the energy storage data acquisition component acquires the (range 26%-30%, calculated = 28%) of 10 energy storage units through a CAN bus, (all 92%-95%, calculated = 92%) and the rated power = 30 kW, calculated = 10 x 28% x 30 = 84 kW; the charging data acquisition component acquires the (total demand = 180 kW) of 20 charging piles through an OCPP 1.6 protocol, and identifies the number of Level 1 users through an RFID card reading module = 3 (all reporting the expected vehicle time = 30 min, calculated = 3); the photovoltaic data acquisition component acquires the = 150 kW of a 1 MW photovoltaic inverter through a Modbus TCP protocol. In the preprocessing link, the = 105% outlier of 1 energy storage unit due to communication interference is rejected (replaced by the previous cycle 28%), all data is converted into the IEC 61850 format, and the , is compressed by difference, and the output contains , , , , , , , 、 、 standardized data, providing input for subsequent decision-making.

[0090] Further, the step S110 further includes the following sub-steps:

[0091] The step S111 collects the real-time frequency of the power distribution network through the frequency sensor The frequency deviation is calculated 50Hz is the rated frequency of the power grid;

[0092] The real-time power of the 10kV bus is collected through the voltage and current sensor The load rate is calculated ;

[0093] The instruction power is obtained through the power grid dispatching interface The instruction power discharge is positive and charge is negative;

[0094] Step S112, through the CAN bus, collect the and the , calculate the , the ; Collect the of each energy storage unit, calculate the total available power ;

[0095] Step S113, through the charging protocol interface, collect the of M charging piles, calculate the total charging demand ; Through the user identification component, identify the user priority, and statistically obtain the number of Level1 users ;

[0096] Based on the user-reported estimated driving time , statistically obtain the number of emergency users , the number of emergency users corresponds to Level1 users;

[0097] Step S114, through the photovoltaic communication interface, collect the real-time output of the photovoltaic inverter The photovoltaic communication interface supports Modbus TCP protocol, and the data collection delay is not more than 30ms.

[0098] ​​​​Among them, "frequency sensor" refers to the detection equipment used to collect the real-time operating frequency of the distribution network, with an accuracy of ±0.01Hz and a sampling frequency of 500Hz to capture instantaneous frequency fluctuations; "voltage and current sensor" refers to the device installed at the 10kV bus to collect the real-time voltage and current of the bus, and then calculate the real-time power P_bus of the bus, with a measurement error ≤0.5%; "CAN bus" refers to the communication bus connecting the local controller of each distributed energy storage unit with the data acquisition module, with a baud rate of 500kbps to ensure data transmission delay ≤10ms to avoid lag in energy storage status data; "charging protocol interface" refers to the communication interface supporting the OCPP 1.6 protocol, used to establish data interaction with charging piles and obtain the charging power demand of each charging pile in real time; "user identification component" refers to the component integrating an RFID card reader module and user APP data receiving function, which identifies user priority (Level 1 / 2) by reading user cards or receiving information reported by the APP; "photovoltaic communication interface" refers to the component supporting Modbus The TCP protocol interface is used to obtain real-time power output data from the photovoltaic inverter, with a data acquisition delay of ≤30ms, ensuring that the photovoltaic power output data can be used in power balance calculations in a timely manner.

[0099] In one possible implementation, power grid data acquisition, in addition to being obtained through the power grid dispatch interface, also involves... In addition, edge computing nodes can parse data from the power distribution network's SCADA system to obtain more detailed regional load distribution information (such as real-time load of each factory building within the park), providing a more accurate grid-side reference for subsequent power allocation; in the energy storage data acquisition stage, besides collecting... , In addition, it can collect the real-time temperature of each energy storage unit (accuracy ±1℃) to determine whether there is a risk of overheating in the energy storage and further enhance the safety protection of energy storage.

[0100] Step S200: Based on the standardized data, first, a hard constraint check of energy storage safety is performed to determine the safety status. Then, the scenario threshold is adaptively adjusted according to the grid load rate and the user's charging urgency to output the scenario type and basic charging power. Subsequently, a preliminary power command is generated through a lightweight reinforcement learning algorithm. Finally, the compliance of the preliminary power command is judged, and the final power command is output.

[0101] In the embodiment of the present application, step S200 is the "decision center" of the entire control method, and its purpose is to dynamically balance the relationship among "grid auxiliary demand, user charging demand, and energy storage protection demand" based on the standardized data output in step S100, to avoid functional conflicts (such as grid peak shaving and user charging competing for energy storage resources), and to finally generate safe, compliant, and executable power instructions. The core effect is to ensure that the grid response meets the standard, the user charging is guaranteed, and the energy storage is not damaged. Among them, "safe state" refers to the energy storage operating state determined after passing the energy storage safety hard constraint check, including "energy storage protection priority state", "limited participation state" and "normal participation state"; "grid load rate" refers to the ratio of the current real-time power of the distribution network to the bus rated capacity, which is the core index for judging the operation pressure of the grid; "user charging urgency" refers to the urgency of the user's charging demand determined according to the user's reported estimated vehicle time or user priority; "scene type" refers to the grid operation scene determined according to the relationship between the frequency deviation and the corresponding threshold, mainly including "grid emergency scene" and "grid normal scene"; "basic charging power" refers to the minimum charging power guarantee value set for Level1 users (high-priority users) to ensure the basic charging demand of emergency users; "preliminary power instruction" refers to the power distribution instruction generated by the lightweight reinforcement learning algorithm, which has not been verified for compliance; "final power instruction" refers to the executable power instruction confirmed to be safe and compliant after compliance judgment.

[0102] In addition, it should be noted that step S200 first defines the safe boundary of energy storage participation through the energy storage safety hard constraint check to avoid decision damage to the energy storage; then performs scene adaptation according to the actual situation of the grid and the user to ensure that the decision fits the current operating environment; then generates a preliminary instruction through a lightweight reinforcement learning algorithm to achieve multi-objective optimization; finally, through compliance judgment, the unsafe and non-compliant instructions are excluded to ensure that the instruction can be implemented.

[0103] In a possible implementation, the lightweight reinforcement learning algorithm adopts the DQN-Lite algorithm, the state space of which only contains 6-dimensional core parameters (frequency deviation, grid instruction power, average state of charge of energy storage, minimum state of health of energy storage, number of Level1 users, and total charging demand), the reward function adopts a three-section type (grid standard reward, charging guarantee reward, and energy storage protection penalty), and the experience replay pool is divided into "grid emergency sub-pool", "grid normal sub-pool", and "energy storage protection sub-pool" according to the scene to ensure balanced learning of each scene, and the decision delay is controlled within 35ms.

[0104] The step S200 further includes the following sub-steps:

[0105] Step S210, if the then switch to the energy storage protection priority mode, prohibit all energy storage from participating in grid assistance, and only allocate basic charging power to Level 1 users , and the total charging power , 0.2 is the maximum allowed discharge depth;

[0106] If , allow the energy storage to participate in grid assistance, but the discharge power , 0.3 is the maximum allowed discharge depth, and basic charging power is allocated to Level 1 users , suspend Level 2 user charging;

[0107] If , continue to perform step S220;

[0108] Step S220, adjust the emergency threshold and normal threshold of the frequency deviation based on the load rate , output a grid emergency scenario or a grid normal scenario;

[0109] Adjust the basic charging power of Level 1 users according to the emergency user proportion ;

[0110] Step S230, construct a state space based on the 6-dimensional core parameters in the standardized data, the 6-dimensional core parameters including , , , , , ;

[0111] Calculate the reward value through a three-section reward function , wherein is the grid compliance reward, the grid responsiveness , , , , , , is the charging guarantee reward, wherein the Level 1 interruption rate , , , , , , is the energy storage protection penalty, wherein , , , , , ;​

[0112] The algorithm model is trained using a scenario-layered experience replay mechanism, and an initial power command is output. The initial power command includes the command power under power grid emergency scenarios. With initial charging power and preliminary grid auxiliary power under normal grid scenarios ;

[0113] Step S240: Check whether the preliminary power command meets the energy storage safety constraints and grid standard constraints, wherein the energy storage safety constraints are the total commanded power. The power grid standard constraint is the power grid command power. , This represents the maximum allowable response power of the power grid.

[0114] If satisfied, output the final power command;

[0115] If the conditions are not met, return to step S230 to readjust the reward function weights and generate a new preliminary power command. The number of retries is ≤3. If the conditions are still not met after 3 retries, the energy storage protection priority mode is triggered.

[0116] Among them, "hard constraint check for energy storage safety" refers to the check based on... and This involves determining whether energy storage meets the safety conditions for participating in grid auxiliary power or meeting charging needs, and avoiding over-discharge or excessive degradation of energy storage health; "scenario threshold adaptation" refers to adjusting based on grid load factor. Adjusting frequency deviation Emergency / normal thresholds, based on the proportion of emergency users. Adjusting the base charging power for Level 1 users allows decisions to adapt to real-time grid and user status; the "lightweight reinforcement learning algorithm" refers to the DQN-Lite algorithm, which reduces the state space from the traditional 10-15 dimensions to 6 dimensions of core parameters. , , , , , The system employs scenario-based experience replay (with a grid emergency / normal / energy storage protection sub-pool ratio of 3:5:2) to improve decision-making efficiency (delay ≤ 35ms) and accuracy (≥ 98%). "Compliance judgment" refers to checking whether the initial power command meets the requirement of "total command power ≤..." (Energy storage security constraints) and "Grid command power ≤ (Power grid standard constraints,) The maximum allowable response power of the power grid is pre-configured by the dispatch center to avoid commands exceeding the hardware or power grid capacity.

[0117] In a possible implementation, the energy storage safety hard constraint check is based on In addition to The real-time discharge depth (D) of each energy storage unit can also be combined If any energy storage unit (corresponding <20%), the discharge power of the energy storage unit is individually limited (≤20% of the rated power), rather than the overall limit of all energy storages, improving the utilization of energy storage resources; In the reward function of the lightweight reinforcement learning algorithm, a "photovoltaic consumption reward R4" can be introduced - if The utilization rate (P) The actual consumption ≥90%, R4=3, further encouraging the system to preferentially use photovoltaic energy and reduce energy storage consumption.

[0118] For example, in a possible implementation, step S210, =22%≥20%, =91%, enter the scene adaptation; step S220, =85% (60%≤ ≤90%), =+0.4Hz (exceeding the normal threshold value ±0.3Hz, determined as an emergency scenario of the power grid), ( >50%), =15kW / pole; step S230, based on the 6-dimensional core parameters, the state space is constructed, and the three-section reward function is calculated =10 ( ≥95%) +8 ( =0) +0 ( ≥25%) =18, output the preliminary instruction: =180kW, =5×15=75kW; step S240, =200kW, the total power of the instruction =180+75=255kW, =260kW (satisfying the constraint), output the final instruction: =180kW, =75kW, complete the dynamic priority decision.

[0119] Further, the step S220 further includes the following sub-steps:

[0120] Step S221, if , the emergency threshold value of is set to ±0.3Hz, and the normal threshold value is set to ±0.2Hz, and the response trigger delay is ≤20ms;

[0121] If , the emergency threshold of the power grid is set to ±0.5 Hz, the normal threshold is set to ±0.3 Hz, and the response triggering delay is less than or equal to 30 ms;

[0122] , the emergency threshold of the power grid is set to ±0.8 Hz, the normal threshold is set to ±0.5 Hz, and the response triggering delay is less than or equal to 50 ms;

[0123] In step S222, if , the basic charging power of the power grid is , the real-time output is preferentially allocated to the basic charging power ;

[0124] , the basic charging power of the power grid is , the basic charging power of the power grid is , and the remaining energy storage capacity is used to supply power in combination with the energy storage;

[0125] , the basic charging power of the power grid is , and the remaining energy storage capacity is used to supply power in combination with the power grid. , the emergency threshold of the power grid refers to a frequency deviation threshold for determining that the power grid enters an emergency scenario, and the threshold is exceeded to preferentially guarantee power grid assistance; the normal threshold of the power grid refers to a frequency deviation range in which the power grid is in a stable operation state, and the charging demand can be preferentially met at this time; the emergency user proportion refers to a ratio of a number of Level 1 users with a predicted driving time of less than or equal to 1 hour to a total number of Level 1 users, and directly reflects the charging emergency degree of the users.

[0126]

[0127] In a possible implementation, in addition to being based on , the power grid side threshold adjustment can also be combined with a “peak regulation level” (such as a first-level peak regulation and a second-level peak regulation) issued by a power grid dispatching center. When the first-level peak regulation (the power grid is extremely heavily loaded), the emergency threshold can be further tightened to ±0.2 Hz, the response triggering delay is less than or equal to 15 ms, and the emergency response speed of the power grid is improved; in addition to being based on , the user side basic charging power adjustment can also be combined with a user charging mode (such as fast charging and slow charging). When the emergency user selects fast charging, , the basic charging power can be increased to 20 kW per pile, and the charging time of the emergency user is further shortened.

[0128] For example, in a possible implementation, in step S221, = 92% (when > 90%), ​​​​​​​​Emergency threshold is set to ± 0.3 Hz, normal threshold is set to ± 0.2 Hz, response trigger delay is set to 18 ms (≤ 20 ms), real-time = + 0.32 Hz (exceeds emergency threshold, determined as grid emergency scenario); step S222, number of Level 1 users = 4 units, of which 3 = 35 min (emergency), 1 unit = 120 min (non-emergency), = 75% (> 50%), = 15 kW / pole, and preferentially allocate photovoltaic output = 120 kW (current photovoltaic output is sufficient), avoid direct consumption of energy storage, through the adjustment, both ensure quick response in grid emergency and guarantee emergency user charging demand.

[0129] Step S300, calculate total power demand according to the final power instruction, allocate energy storage power and adjust charging power according to scenario type, and output actual operation data after execution.

[0130] In the embodiments of the present application, the "total power demand" refers to the total power currently required by the system calculated according to the final power instruction, which includes the sum of grid auxiliary required power and user charging required power; the "charging power adjustment" refers to adjusting the output power of each charging pile according to the charging power determined in the final power instruction to ensure that the charging demand meets the instruction; the "actual operation data" refers to the actual operation parameters collected from the energy storage unit and the charging pile after the execution of power allocation and adjustment, including but not limited to the actual response power of the energy storage unit, the actual charging power of the charging pile, the equipment operating temperature, etc.

[0131] In addition, it should be noted that step S300 first calculates the total power demand according to the final power instruction to determine the current power consumption or total output of the system; then determines the priority according to the scenario type (grid emergency scenario prioritizes grid instruction, grid normal scenario prioritizes charging demand); finally, the power is allocated to each energy storage unit, and the charging pile power is adjusted, while the actual operation data is collected. The execution subject is a distributed execution module in the system, which includes energy storage control components (such as bidirectional inverters, local power controllers) and charging control components (such as charging pile power regulation units) to ensure accurate execution of instructions.

[0132] In a possible implementation, the energy storage power allocation adopts a "capacity-health degree weighting algorithm", that is, the power allocation coefficient is equal to the product of the state of charge and the state of health of a single energy storage unit, divided by the sum of the products of the state of charge and the state of health of all energy storage units. This algorithm can ensure that energy storage units with high state of charge and good state of health bear more power, avoid overloading of single units, and prolong the overall energy storage life.

[0133] wherein the step S300 further comprises the following sub-steps:

[0134] Step S310, if it is an emergency scenario of the power grid, then the total discharge power requirement , is the final grid instruction power is the final charging power

[0135] If , then the power distribution coefficient is used to distribute the power of each energy storage unit , wherein i = 1, 2, …, N

[0136] If , then the charging power reduction ratio is calculated, and the actual charging power is obtained

[0137] If it is a normal scenario of the power grid, then the total power consumption requirement , is the final grid auxiliary power

[0138] If , then the is preferentially met, and the remaining power is used as the actual grid auxiliary power ; if , then the grid auxiliary power reduction ratio is calculated, and the actual grid auxiliary power is obtained

[0139] Step S320, the energy storage control component sends the power of each energy storage unit or to each energy storage unit, so that the energy storage inverter adjusts the power according to the PWM control signal and outputs the actual grid response power ; the charging control component sends to each charging pile, so that the charging pile adjusts the output current and outputs the actual charging power ;

[0140] If an overcurrent or overtemperature anomaly is detected, then an emergency stop switch is triggered, power distribution is suspended, and an emergency mode is switched to, wherein the emergency mode only guarantees Level 1 user 5kW / pile charging, wherein the overcurrent is a current > 1.2x rated current, and the overtemperature is a temperature > 65℃.

[0141] wherein “power balance calculation” refers to calculating the actual power distribution value of each device according to the final power instruction and the total available power of the energy storage and the photovoltaic output, so as to avoid the total power exceeding the system carrying capacity; and “power distribution coefficient” refers to the power distribution coefficient of each device "Refers to based on each energy storage unit" and The calculated weighting coefficients ensure balanced load across all energy storage units (power deviation ≤ 5%), preventing overload of any single unit. A "bidirectional inverter" refers to a device that converts stored DC power into AC power (discharging) or vice versa (charging), with an efficiency ≥ 98.5% and supports PWM (pulse width modulation) control for continuously adjustable power. An "emergency stop switch" is a protection device installed on both the energy storage inverter and the charging pile, with a response time ≤ 100ms. It triggers when overcurrent (current > 1.2 times rated current) or overtemperature (temperature > 65℃) is detected, cutting off power output to protect the equipment.

[0142] In one possible implementation, the power balance calculation step, in addition to considering and In addition, "diesel generator backup power" (such as the 2MW diesel generator configured in the park) can be introduced. When the power grid is in an emergency and the energy storage + photovoltaic power is insufficient, the diesel generator can be started to supplement the power, further improving the grid's auxiliary response capability; in the distributed execution link, in addition to feedback and In addition, it can also provide feedback on the real-time operating temperature and current of each device, providing more comprehensive device status data for subsequent feedback and correction steps.

[0143] For example, in one feasible implementation, step S310, in a power grid emergency scenario, =250kW, =60kW (3 Level 1 users, =20kW / pile), total demand =310kW, =280kW, =50kW, total available =330kW≥310kW, according to Allocate 250kW of grid power ( Range 0.09-0.11 (load balancing), 60kW charging power; Step S320, the energy storage control component sends power commands (23-27kW / unit) to 10 energy storage units, the inverter controls the output power through PWM, and feedback... =250kW; The charging control component sends a 20kW / unit command to each of the three charging piles, and provides feedback. =60kW; During execution, the temperature of one inverter rose to 68℃ (>65℃), triggering the emergency stop switch. This energy storage unit stopped operating, and the system automatically distributed its 25kW power to the other 9 energy storage units. (Recalculate) to ensure the total power remains unchanged, and complete the power allocation execution.

[0144] Step S400, calculate a deviation value based on the actual operation data and the final power instruction, if the deviation value exceeds a preset threshold, adjust the energy storage power distribution parameter and push the state information to the power grid, the user and the next round of decision-making link to form a closed-loop control.

[0145] In the embodiments of the present application, step S400 is the core link to realize the "continuous optimization" of the system. The purpose is to find the execution deviation by comparing the difference between the actual operation data and the final power instruction, and to adjust it, at the same time, to push the state information to the relevant parties, and finally to form a "collection-decision-execution-feedback-optimization" closed-loop control. The core effect is to improve the stability and accuracy of the system operation, and to reduce the deviation of subsequent decision-making. Among them, the "deviation value" refers to the difference between the actual operation data and the final power instruction, which is usually expressed in percentage form to quantify the size of the execution deviation; the "preset threshold" refers to the upper limit of the allowed deviation range set by the system in advance. If the deviation value exceeds the threshold, it needs to be corrected; the "energy storage power distribution parameter" refers to the parameter used in the process of energy storage power distribution, such as power distribution coefficient, charging power reduction ratio, etc.; the "state information" refers to the current operation state data of the system, including the grid response, the user charging progress, the remaining capacity of the energy storage, etc.; the "next round of decision-making link" refers to the process of executing step S200 (dynamic priority decision step) next time; the "closed-loop control" refers to the transmission of the execution deviation information to the next round of decision-making through feedback correction, so that the next round of decision-making can be adjusted accordingly, forming a cyclic optimization control mode.

[0146] In addition, it should be noted that step S400 first calculates various deviation values (such as grid response deviation, charging power deviation) based on the actual operation data and the final power instruction; then compares the deviation value with the preset threshold to determine whether it needs to be corrected; if it needs to be corrected, adjust the energy storage power distribution parameter (such as adjust the power distribution coefficient), and push the state information to the grid dispatching center, the user and the next round of decision-making link; if it does not need to be corrected, only push the state information to ensure that the relevant parties understand the system operation state. The execution subject is the feedback module in the system, which realizes the deviation calculation, correction and information pushing through the interaction with the distributed execution module and the central decision-making module.

[0147] In one possible implementation, the preset threshold is set to 10%, that is, if the grid response deviation or the charging power deviation exceeds 10%, the correction is triggered; the state information pushing adopts a multi-channel mode: uploading the actual response power and the remaining capacity of the energy storage to the grid dispatching center, pushing the charging progress and the cost to the user, and pushing the equipment health report (such as the SOH attenuation trend of the energy storage) to the operation and maintenance end to ensure that different subjects obtain the required information.

[0148] Among them, the step S400 further includes the following sub-steps:

[0149] Step S410: Calculate the power grid response deviation. ,like but ;

[0150] Calculate charging power deviation ,like but ;

[0151] Calculate photovoltaic output deviation The To contribute to the actual development of photovoltaics;

[0152] Step S420, if Then improve energy storage unit The coefficient is 10%, that is And return to step S200 for feedback. Correction factor;

[0153] like and Then from An additional 5kW / pile will be allocated to Level 1 users, and feedback will be returned to step S300. Fluctuation warning and reservation As backup power;

[0154] Upload to the power grid dispatch center With remaining total available power Push charging progress to users ,cost And abnormal alerts, push equipment health reports to the operation and maintenance terminal, the equipment health reports contain Decline trend Inverter temperature profile During the charging progress This is the amount of charge already applied. To meet the user's power demand, It is a time-of-use electricity price.

[0155] in," "Correction factor" refers to the power allocation factor for high SOC energy storage units when the grid response deviation exceeds the limit. Adjustments (such as a 10% increase) will be made to supplement the grid response power; "Fluctuation warning" refers to sending an early warning to the central decision-making module when the photovoltaic output deviation exceeds 15%, prompting the next round of decision-making to reserve more energy storage backup power to cope with photovoltaic fluctuations; "Remaining total available power" refers to the remaining capacity after subtracting the actual grid response power and actual charging power from the total available energy storage power. This is used to provide feedback to the power grid regarding subsequent response capabilities.

[0156] In one possible implementation, the deviation calculation step, in addition to calculating... , , In addition, it can also calculate the "energy storage power distribution deviation". ( ),in This refers to the actual power output of the i-th distributed energy storage unit (i=1,2,…,N, where N is the total number of energy storage units) in step S300. This refers to the power allocation command value issued to the i-th distributed energy storage unit, which is the target value calculated based on the final power command output in step S200. If a certain energy storage unit... If the failure rate is greater than 10%, the unit is considered to be potentially faulty, and a maintenance reminder is sent to the operations and maintenance team; the execution process is corrected, including adjustments... In addition to reserving backup power, the weight of the reward function of the lightweight algorithm can be dynamically adjusted. For example, when photovoltaic fluctuations are large, the weight of R4 (photovoltaic grid connection reward) can be increased to further optimize subsequent decisions.

[0157] For example, in one feasible implementation, step S410, =200kW, =185kW, (≤10%, qualified); =80kW, =72kW, =10% (critical pass); =150kW, =120kW, =20% (>15%, unacceptable); Step S420, because... >15%, from An additional 5kW / pile will be allocated from the 300kW to 2 Level 1 users. Adjusted to 80kW), feedback was sent to the central decision-making module. Fluctuation warning, reserve 10% × 300 = 30kW of standby power; in the status push phase, upload data to the power grid. =185kW =300-185-80=35kW; Push charging progress (80% charged) and cost (80kW×1.2 yuan / kWh=96 yuan) to the user; Push equipment health report to the maintenance department ( The decline trend is 0.5% / month (normal). Feedback and correction have been completed, providing an optimization basis for the next round of decision-making.

[0158] In the embodiment, by constructing a closed-loop control architecture of data acquisition, dynamic decision, power execution and feedback correction, balancing the demands of power grid, charging and energy storage, the effect of improving the reliability of power grid interaction and the charging experience of users is achieved; by lightweight reinforcement learning algorithm and scenario threshold self-adaptation, the decision efficiency and adaptability are optimized, the effect of reducing system response delay and enhancing multi-device compatibility is achieved; by energy storage safety hard constraint checking and hierarchical protection mechanism, over-discharge and health decay of energy storage are avoided, the effects of prolonging the life cycle of energy storage and reducing the replacement cost of devices are achieved.

[0159] On the basis of the above-mentioned embodiments, as shown in Figure 2 The application also provides a multi-functional collaborative control system for distributed energy storage charging, for supporting the multi-functional collaborative control method for distributed energy storage charging of the above-mentioned embodiments, the multi-functional collaborative control system for distributed energy storage charging comprises:

[0160] A data acquisition module is configured to acquire original data of a power grid side, an energy storage side, a charging side and a photovoltaic side, perform outlier rejection, protocol conversion and data compression processing on the original data, and output standardized data.

[0161] A central decision module is configured to, based on the standardized data, first perform energy storage safety hard constraint checking to determine a safety state, then perform scenario threshold self-adaptation according to a power grid load rate and a user charging urgency to output a scenario type and a basic charging power, subsequently generate a preliminary power instruction through a lightweight reinforcement learning algorithm, and finally perform compliance judgment on the preliminary power instruction to output a final power instruction.

[0162] A distributed execution module is configured to calculate a total power demand according to the final power instruction, perform energy storage power distribution and charging power adjustment according to the scenario type, and output actual operation data after execution.

[0163] A feedback module is configured to calculate a deviation value based on the actual operation data and the final power instruction, and if the deviation value exceeds a preset threshold, adjust an energy storage power distribution parameter and push state information to a power grid, a user and a next round of decision link to form a closed-loop control.

[0164] In the embodiment, by constructing a closed-loop control architecture of data acquisition, dynamic decision, power execution and feedback correction, balancing the demands of power grid, charging and energy storage, the effect of improving the reliability of power grid interaction and the charging experience of users is achieved; by lightweight reinforcement learning algorithm and scenario threshold self-adaptation, the decision efficiency and adaptability are optimized, the effect of reducing system response delay and enhancing multi-device compatibility is achieved; by energy storage safety hard constraint checking and hierarchical protection mechanism, over-discharge and health decay of energy storage are avoided, the effects of prolonging the life cycle of energy storage and reducing the replacement cost of devices are achieved.

[0165] Further, the distributed energy storage charging multifunctional collaborative control system can run the distributed energy storage charging multifunctional collaborative control method described above, and specific implementation can be referred to the method embodiments, which will not be described here.

[0166] On the basis of the above embodiments, as Figure 3 The electronic device includes:

[0167] At least one processor 22, at least one memory 21, a communication interface 23 and a communication bus 24, the processor 22 is in communication connection with the memory 21;

[0168] In this embodiment, the memory 21 can be implemented in any appropriate manner, for example: the memory 21 can be a read-only memory, a mechanical hard disk, a solid state disk or a U disk, etc.; the memory 21 is used to store executable instructions executed by at least one processor;

[0169] In this embodiment, the processor 22 can be implemented in any appropriate manner, for example, the processor 22 can take the form of, for example, a microprocessor or a processor and a computer readable medium storing computer readable program code (such as software or firmware) executable by the (micro) processor, logic gates, switches, application specific integrated circuits (ASIC), programmable logic controllers and embedded microcontrollers, etc.; the processor is used to execute the executable instructions to realize the distributed energy storage charging multifunctional collaborative control method as described above.

[0170] On the basis of the above embodiments, the present application further provides a computer readable storage medium, the computer readable storage medium has a computer program stored therein, the computer program is executed by the processor to realize the distributed energy storage charging multifunctional collaborative control method as described above.

[0171] Those of ordinary skill in the art can realize that the modules and method steps of the examples described in combination with the embodiments disclosed herein can be realized in electronic hardware, or a combination of computer software and electronic hardware. Whether the functions are executed in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of the present application.

[0172] Those skilled in the art can clearly understand that, for the convenience and brevity of the description, the specific working process of the above-described apparatus, device and module can refer to the corresponding process in the foregoing method embodiments, which will not be described here.

[0173] In several embodiments provided in the present application, it should be understood that the disclosed devices, apparatuses and methods can be implemented in other manners. For example, the embodiments of the apparatus described above are merely schematic. For example, the division of the modules is merely logical function division. There can be another division manner for the actual implementation. For example, a plurality of modules or a unit can be combined or integrated into another apparatus, or some features can be ignored or not executed. In addition, the displayed or discussed mutual couplings or direct couplings or communication connections can be indirect couplings or communication connections through some interfaces, apparatuses or units, and can be in electrical, mechanical or other forms.

[0174] The modules illustrated as separated components can or can not be physically separated, and the components illustrated as modules can or can not be physical units, i.e., can be located in one place, or can be distributed on a plurality of network units. Some or all of the units can be selected according to actual needs to achieve the purposes of the embodiments.

[0175] In addition, each functional module in each embodiment of the present application can be integrated into a processing module, or each module can be physically present alone, or two or more modules can be integrated into one module.

[0176] If the functions are realized in the form of software function modules and sold or used as independent products, they can be stored in a computer readable storage medium. Based on this understanding, the technical solutions of the present application essentially or the parts that make contributions to the prior art or parts of the technical solutions can be embodied in the form of a software product. The computer software product is stored in a storage medium and includes a plurality of instructions for causing a computer device (which can be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present application. The aforementioned storage medium includes: a U disk, a mobile hard disk, a read-only memory server, a random access memory server, a magnetic disk or an optical disk, and various media that can store program instructions.

[0177] It should be noted that the above enumeration is only specific embodiments of the present application, and obviously the present application is not limited to the above embodiments, and there are many similar changes. All modifications directly derived or thought by those skilled in the art from the content disclosed by the present application should be included in the protection scope of the present application.

[0178] The above are only the preferred embodiments of the present application, and are not used to limit the protection scope of the present application. Any modification, equivalent replacement, improvement, etc. made within the spirit and principles of the present application should be included in the protection scope of the present application.

Claims

1. A multi-functional collaborative control method for distributed energy storage charging, characterized in that, Includes the following steps: Step S100: Collect raw data from the grid side, energy storage side, charging side and photovoltaic side, perform outlier removal, protocol conversion and data compression on the raw data, and output standardized data; Step S200: Based on the standardized data, first perform a hard constraint check on energy storage safety to determine the safety status, then perform scenario threshold adaptation based on grid load rate and user charging urgency to output scenario type and basic charging power, then generate preliminary power command through lightweight reinforcement learning algorithm, and finally perform compliance judgment on the preliminary power command to output final power command. Step S300: Calculate the total power demand according to the final power command, allocate energy storage power and adjust charging power according to the scenario type, and output the actual operating data after execution; Step S400: Calculate the deviation value based on the actual operating data and the final power command. If the deviation value exceeds the preset threshold, adjust the energy storage power allocation parameters and push the status information to the power grid, users and the next round of decision-making to form a closed-loop control. Step S200 further includes the following sub-steps: Step S210, if In this mode, the system switches to energy storage protection priority mode, prohibiting all energy storage from participating in grid auxiliary operations and allocating basic charging power only to Level 1 users. And total charging power 0.2 represents the maximum permissible depth of discharge, where, The unit is in a charged state. The average state of charge, Rated power; like This allows energy storage to participate in grid auxiliary operations, but the discharge power... 0.3 represents the maximum permissible depth of discharge, allocating basic charging power to Level 1 users. Charging for Level 2 users is suspended. like If so, continue with step S220; Step S220, based on the load factor Adjusting frequency deviation The emergency threshold and normal threshold are used to output either the emergency scenario or the normal scenario of the power grid. Based on the proportion of emergency users Adjust the base charging power for Level 1 users ,in, Number of emergency users; Step S230: Construct a state space based on the 6-dimensional core parameters in the standardized data. The 6-dimensional core parameters include... , , , , , ,in, For distribution network frequency deviation, For command power, For the minimum healthy state, For Level 1 users, Total charging demand; The reward value is calculated using a three-stage reward function. ,in Incentives for grid compliance and grid responsiveness hour , hour , hour , As a reward for charging protection, the Level 1 interruption rate is included. hour , hour , hour , Penalties for energy storage protection, among which hour , hour , hour ; The algorithm model is trained using a scenario-layered experience replay mechanism, and an initial power command is output. The initial power command includes the command power under power grid emergency scenarios. With initial charging power and preliminary grid auxiliary power under normal grid scenarios ; Step S240: Check whether the preliminary power command meets the energy storage safety constraints and grid standard constraints, wherein the energy storage safety constraints are the total commanded power. , The total available power is defined as the grid standard constraint, which is the grid command power. , This is the maximum allowable response power of the power grid; If satisfied, output the final power command; If the conditions are not met, return to step S230 to readjust the reward function weights and generate a new preliminary power command. The number of retries is ≤3. If the conditions are still not met after 3 retries, the energy storage protection priority mode is triggered.

2. The multi-functional collaborative control method for distributed energy storage charging according to claim 1, characterized in that, Step S100 further includes the following sub-steps: Step S110: Collect the frequency deviation on the power grid side using the power grid data acquisition component. Load factor and command power ; The state of charge of each energy storage unit on the energy storage side is collected by the energy storage data acquisition component. Health status and rated power Then the average state of charge is calculated. Minimum health status and total available power ,in, and In this context, i = 1, 2, ..., N, where N is the total number of energy storage units; The charging data acquisition component collects the charging demand of each charging pile on the charging side. Level 1 user count and number of emergency users The total charging demand is calculated. ,in, In the context, j = 1, 2, ..., M, where M is the total number of charging stations; Real-time power output from the photovoltaic side is collected through photovoltaic data acquisition components. ; Step S120: Use the 3σ criterion to remove data from the original data that exceeds a reasonable range. The data exceeding a reasonable range includes... , or , or , For the rated power of photovoltaic power, the data after elimination is replaced with data from the previous period; Convert the data in the original data of different protocol formats to the IEC 61850 standard format; , , and Differential compression is used, with a compression ratio of no less than 1:5; Step S130: Integrate the data processed in step S120 and output the data containing... , , , , , , , , , The standardized data mentioned above.

3. The multi-functional collaborative control method for distributed energy storage charging according to claim 2, characterized in that, Step S110 further includes the following sub-steps: Step S111: Collect the real-time frequency of the power distribution network using a frequency sensor. The frequency deviation was calculated. 50Hz is the rated frequency of the power grid; Real-time power of the 10kV bus is collected using voltage and current sensors. Combined with the rated capacity of the busbar The load factor was calculated. ; The commanded power is obtained through the power grid dispatching interface. The command power Discharging is positive, charging is negative; Step S112: Collect the data from N distributed energy storage units via the CAN bus. With the The calculation yielded the The above ; The data collected from each energy storage unit The total available power is calculated. ; Step S113: Collect the data from the M charging piles via the charging protocol interface. The total charging demand is calculated. ; The user priority is identified by the user identification component, and the number of Level 1 users is obtained by counting. ; Based on the user's reported estimated car usage time The number of emergency users was obtained by statistics. The number of emergency users correspond Level 1 users; Step S114: Collect the real-time output of the photovoltaic inverter through the photovoltaic communication interface. The photovoltaic communication interface supports the Modbus TCP protocol, and the data acquisition delay does not exceed 30ms.

4. The multi-functional collaborative control method for distributed energy storage charging according to claim 3, characterized in that, Step S220 further includes the following sub-steps: Step S221, if ,but The emergency threshold is set to ±0.3Hz, the normal threshold is set to ±0.2Hz, and the response trigger delay is ≤20ms; like ,but The emergency threshold is set to ±0.5Hz, the normal threshold is set to ±0.3Hz, and the response trigger delay is ≤30ms; like ,but The emergency threshold is set to ±0.8Hz, the normal threshold is set to ±0.5Hz, and the response trigger delay is ≤50ms; Step S222, if The basic charging power Prioritize the allocation of the real-time output ; like The basic charging power ,use Combined supply with energy storage; like The basic charging power The remaining energy storage capacity participates in grid auxiliary operations.

5. The multi-functional collaborative control method for distributed energy storage charging according to claim 1, characterized in that, Step S300 further includes the following sub-steps: Step S310: If it is a power grid emergency scenario, then the total discharge power requirement is... , For the final grid command power, This refers to the final charging power. like Then according to the power allocation coefficient Power allocation for each energy storage unit ,in i=1,2,…,N; like Then calculate the charging power reduction ratio. To obtain the actual charging power ; If it is a normal power grid scenario, the total power consumption demand is... , As auxiliary power to the final power grid; like Then priority will be given to satisfying. The surplus power serves as auxiliary power for the actual power grid. ;like Then calculate the reduction ratio of auxiliary power in the power grid. To obtain the actual grid auxiliary power ; Step S320: The power of each energy storage unit is sent to each energy storage unit through the energy storage control component. or The drive energy storage inverter adjusts its power according to the PWM control signal, outputting the actual grid response power. ; The charging control component sends commands to each charging station. Adjust the charging pile's output current to output the actual charging power. ; If an overcurrent or overtemperature anomaly is detected, the emergency stop switch is triggered, power distribution is suspended, and the system switches to emergency mode. The emergency mode only guarantees 5kW / pile charging for Level 1 users. The overcurrent is defined as current > 1.2 × rated current, and the overtemperature is defined as temperature > 65℃.

6. The multi-functional collaborative control method for distributed energy storage charging according to claim 5, characterized in that, Step S400 further includes the following sub-steps: Step S410: Calculate the power grid response deviation. ,like but ; Calculate charging power deviation ,like but ; Calculate photovoltaic output deviation The To contribute to the actual development of photovoltaics; Step S420, if Then improve energy storage unit The coefficient is 10%, that is and return to step S300 for feedback. Correction factor; like and Then from An additional 5kW / pile will be allocated to Level 1 users, and feedback will be returned to step S300. Fluctuation warning and reservation As backup power; Upload to the power grid dispatch center With remaining total available power Push charging progress to users ,cost And abnormal alerts, push equipment health reports to the operation and maintenance terminal, the equipment health reports contain Decline trend Inverter temperature profile During the charging progress This is the amount of charge already applied. To meet user power needs, It is a time-of-use electricity price.

7. A multi-functional collaborative control system for distributed energy storage and charging, characterized in that, include: The data acquisition module is used to collect raw data from the grid side, energy storage side, charging side and photovoltaic side, and to perform outlier removal, protocol conversion and data compression on the raw data to output standardized data. The central decision-making module is used to first perform a hard constraint check on energy storage safety to determine the safety status based on the standardized data, then perform scenario threshold adaptation based on the grid load rate and the user's charging urgency to output the scenario type and basic charging power, then generate a preliminary power command through a lightweight reinforcement learning algorithm, and finally perform compliance judgment on the preliminary power command to output the final power command. The distributed execution module is used to calculate the total power demand according to the final power command, allocate energy storage power and adjust charging power according to the scenario type, and output the actual operation data after execution. The feedback module is used to calculate the deviation value based on the actual operating data and the final power command. If the deviation value exceeds the preset threshold, the energy storage power allocation parameters are adjusted and the status information is pushed to the power grid, users and the next round of decision-making to form a closed-loop control. The central decision-making module is also used for: like In this mode, the system switches to energy storage protection priority mode, prohibiting all energy storage from participating in grid auxiliary operations and allocating basic charging power only to Level 1 users. And total charging power 0.2 represents the maximum permissible depth of discharge, where, The unit is in a charged state. The average state of charge, Rated power; like This allows energy storage to participate in grid auxiliary operations, but the discharge power... 0.3 represents the maximum permissible depth of discharge, allocating basic charging power to Level 1 users. Charging for Level 2 users is suspended. like Then continue based on the load factor Adjusting frequency deviation The emergency threshold and normal threshold are used to output either the emergency scenario or the normal scenario of the power grid. Based on the proportion of emergency users Adjust the base charging power for Level 1 users ,in, Number of emergency users; Then, a state space is constructed based on the 6-dimensional core parameters in the standardized data. These 6-dimensional core parameters include... , , , , , ,in, For distribution network frequency deviation, For command power, For the minimum healthy state, For Level 1 users, Total charging demand; The reward value is calculated using a three-stage reward function. ,in Incentives for grid compliance, grid responsiveness hour , hour , hour , As a reward for charging protection, the Level 1 interruption rate is included. hour , hour , hour , Penalties for energy storage protection, among which hour , hour , hour ; The algorithm model is trained using a scenario-layered experience replay mechanism, and an initial power command is output. The initial power command includes the command power under power grid emergency scenarios. With initial charging power and preliminary grid auxiliary power under normal grid scenarios ; Finally, it is checked whether the preliminary power command meets the energy storage security constraints and grid standard constraints, wherein the energy storage security constraints are the total commanded power. , The total available power is defined as the grid standard constraint, which is the grid command power. , This represents the maximum allowable response power of the power grid. If satisfied, output the final power command; If the conditions are not met, return to step S230 to readjust the reward function weights and generate a new preliminary power command. The number of retries is ≤3. If the conditions are still not met after 3 retries, the energy storage protection priority mode is triggered.

8. An electronic device, characterized in that, The electronic device includes: The processor and the memory are communicatively connected. The memory is used to store at least one executable instruction executed by the processor, the processor being used to execute the executable instruction to implement the multi-functional collaborative control method for distributed energy storage charging as described in any one of claims 1 to 6.

9. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program, which, when executed by a processor, implements the multi-functional collaborative control method for distributed energy storage charging as described in any one of claims 1 to 6.

Images