Method and device for controlling multi-energy asynchronous cooperative frequency modulation power output and electronic equipment

Abstract

This application discloses a control method, device, and electronic device for asynchronous coordinated frequency regulation output of multiple energy storage systems. The method includes: generating a plant-level frequency regulation target based on the upper-level frequency regulation command and the grid connection point power value; determining the command version number by combining a preset adaptive participation factor table; allocating power to the plant-level frequency regulation target within the executable power range of each energy storage system to obtain a set of candidate sub-commands that are mapped one-to-one with the command version number; subsequently issuing a preparation request and verifying consistency, and issuing a submit or withdraw command to the corresponding energy storage system; calculating the grid connection point power deviation in real time, generating and issuing correction sub-commands to the energy storage system; and finally dynamically updating the adaptive participation factor table for use in the command generation and power allocation of the next frequency regulation target, thereby achieving efficient coordinated frequency regulation. This application can avoid distortion of grid connection point frequency regulation output caused by asynchronous reception, inconsistent directions, or parallel execution of new and old commands, and improve the consistency and traceability of grid connection point frequency regulation execution.

Description

Technical Field

[0001] This application relates to the field of power system technology, and in particular to a control method, device and electronic equipment for multi-energy storage asynchronous coordinated frequency regulation output. Background Technology

[0002] When thermal power units participate in grid frequency regulation, the dispatching side typically issues plant-level power regulation commands. Within a short cycle, the power plant needs to ensure that the power at the grid connection point follows the target change to suppress frequency deviation. To improve response speed and bidirectional regulation capability, thermal power plants often configure multiple energy storage systems in parallel at the grid connection point to participate in primary / secondary frequency regulation. The plant-level coordinator needs to integrate the grid connection point power value and the superior command within each frequency regulation cycle to form a plant-level frequency regulation target, and then decompose the target into various energy storage sub-commands. In practice, multiple energy storage systems often come from different manufacturers, resulting in differences in PCS control cycles, power loop bandwidth, limiting logic, and communication protocols. Simultaneously, the plant's industrial Ethernet and uplink links suffer from latency jitter, packet loss, and acknowledgment delays, causing inconsistencies in the reception and activation times of the same plant-level command at different energy storage sides. The grid-connected power is a composite output from multiple energy storage units and generating units. If the energy storage units are not synchronized or the switching between old and new commands is inconsistent, phenomena such as inconsistency between net output and target, instantaneous superposition or cancellation can easily occur, thus affecting the achievement of plant-level frequency regulation targets and process traceability. Therefore, multi-energy storage parallel frequency regulation not only involves power decomposition, but also the organization of command boundary management, execution set definition, and deviation closed-loop correction under heterogeneous controllers and uncertain communication conditions.

[0003] In related technologies, the plant-level coordinator typically allocates the plant-level frequency regulation target based on information such as the capacity, state of charge (SOC), available power, and health status of each energy storage unit, using a fixed ratio or adaptive participation factor. It then issues power setpoints or incremental commands to each energy storage controller. Each energy storage unit tracks the power in a closed-loop manner locally and reports its execution status. Some schemes collect confirmation receipts or operational status after issuance and perform secondary corrections based on the deviation between the grid connection point power value and the target. The correction amount is then redistributed to each energy storage unit according to the participation factor, aiming to form a closed loop at the plant level. The above methods mostly organize the frequency regulation cycle in a single-stage process of issuance-execution-deviation redistribution, assuming that each energy storage unit can consistently receive and take effect on the same command within the same cycle.

[0004] However, under conditions of multiple manufacturers, multiple control cycles, and communication uncertainty, the following problems are likely to occur: First, the lack of clear instruction version boundaries and execution set definitions leads to some units still operating under old instructions and running in parallel with new instructions, resulting in superposition or cancellation of net output; Second, when units are constrained by amplitude / direction, they may have executable states that are inconsistent with the allocation direction, but these are still included in the net force, making it difficult to attribute grid connection point deviations; Third, deviation corrections often continue to be distributed to all energy storage units, causing units that have not received corrections in time or have not taken effect on schedule to participate in the correction path, which will further amplify the output distortion caused by asynchrony. Summary of the Invention

[0005] This disclosure provides a control method, apparatus, equipment, and storage medium for asynchronous coordinated frequency modulation output of multiple energy storage systems, in order to at least solve the above-mentioned technical problems existing in the prior art.

[0006] According to a first aspect of this application, a control method for multi-energy storage asynchronous cooperative frequency modulation output is provided, the method comprising: Based on the obtained upper-level frequency regulation commands and grid connection point power values ​​corresponding to multiple energy storage systems connected in parallel at thermal power plants, plant-level frequency regulation targets are generated. Based on the plant-level frequency regulation target and the preset adaptive participation factor table corresponding to each energy storage system, the instruction version number is determined. Within the executable power range of each energy storage system, the power of the plant-level frequency regulation target is allocated to obtain a set of candidate sub-instructions that are mapped one-to-one with the instruction version number. The adaptive participation factor table is used to characterize the execution reliability of each energy storage system version. The adaptive participation factor table is a pre-initialized weight table that is iteratively updated with the execution effect of the instruction. The instruction version number and the corresponding set of candidate sub-instructions are sent to each energy storage system to form a preparation request. Obtain the preparation confirmation information set returned by each energy storage system in response to the preparation request, and perform a consistency judgment on the preparation request of each energy storage system based on the preparation confirmation information set. Generate a submit command or cancel command for the corresponding energy storage system according to the judgment result. Send a submission command to the energy storage system that generates the submission command, and send a cancellation command to the energy storage system that generates the cancellation command; Based on the plant-level frequency regulation target and the real-time acquired grid connection point power value, the grid connection point power deviation is calculated; using the adaptive participation factor table as the allocation constraint, the grid connection point power deviation is allocated to the energy storage system that has issued the submission instruction, and a set of correction sub-instructions corresponding to the instruction version number is generated. The set of correction sub-instructions is issued to the energy storage systems within the consensus set; Based on the set of corrected sub-instructions, the version execution record of this instruction, the set of preparation confirmation information, and the change in the power deviation at the grid connection point, calculate and generate an instruction execution effectiveness evaluation quantity. The adaptive participation factor table is dynamically updated based on the effectiveness evaluation metric to obtain the updated adaptive participation factor table, and the updated adaptive participation factor table is applied to the instruction generation and power allocation of the next plant-level frequency regulation target.

[0007] In one possible implementation, generating a plant-level frequency regulation target based on the acquired upper-level frequency regulation commands and grid connection point power values ​​corresponding to multiple parallel energy storage systems in a thermal power plant includes: Receive the frequency modulation command from the superior, extract the command arrival time and the command power target value, and use the command arrival time as the alignment reference for this frequency modulation cycle; Within this frequency modulation cycle, the grid connection point power value is collected, the grid connection point power value is time-aligned according to the alignment reference, and the grid connection point power value is averaged using a preset sampling window to obtain the grid connection point power reference value. A plant-level frequency regulation target is generated based on the commanded power target value and the grid connection point power reference value; wherein, the plant-level frequency regulation target includes: execution direction and amplitude limit; the amplitude limit is constrained by a preset plant-level adjustable power boundary.

[0008] In one possible implementation, based on the plant-level frequency regulation target and a preset adaptive participation factor table corresponding to each energy storage system, an instruction version number is determined. Within the executable power range of each energy storage system, power allocation is performed on the plant-level frequency regulation target to obtain a set of candidate sub-instructions that are mapped one-to-one with the instruction version number, including: The adaptive participation factor table pre-stored in the plant-level coordinator is read to obtain the corresponding entries for each energy storage system; the entries include the energy storage identifier, the current value of the adaptive participation factor, the allowed value boundary, and the reliability parameter used to characterize the reliability of version execution. The instruction boundary of the current frequency modulation cycle is determined based on the plant-level frequency modulation target, and the instruction boundary is combined with the execution direction of the plant-level frequency modulation target and the amplitude limit value to generate an instruction version number; Based on the current value of the adaptive participation factor and the plant-level frequency regulation target, initial power allocation is performed to obtain the initial decomposition command value corresponding to each energy storage system; The initial decomposition command value is limited based on the executable power range corresponding to each energy storage system. The remaining power generated by the limiting is redistributed to the energy storage systems that have not touched the boundaries of the executable power range according to the adaptive participation factor table until the decomposition commands of all energy storage systems meet the executable power range constraints. The energy storage identifier, instruction version number, instruction direction, and decomposed instruction value of each energy storage system are encapsulated into a set of candidate sub-instructions.

[0009] In one possible implementation, the formation preparation request includes: Write the energy storage identifier, instruction version number, corresponding candidate sub-instruction direction, and corresponding candidate sub-instruction value into the preparation request, and configure a preparation timeout threshold for the preparation request. The preparation request is sent to each energy storage system, and a timer based on the preparation timeout threshold is started on the plant-level coordinator side.

[0010] In one possible implementation, the step of obtaining the preparation confirmation information set returned by each energy storage system in response to the preparation request, and performing a consistency judgment on the preparation request of each energy storage system based on the preparation confirmation information set, and generating a submit command or cancel command for the corresponding energy storage system according to the judgment result, includes: Within the preparation timeout threshold, receive preparation confirmation information returned by each energy storage system and summarize it into a preparation confirmation information set according to the energy storage identifier; Extract the receiving instruction version number and preparation confirmation status of each energy storage system from the preparation confirmation information set, compare the receiving instruction version number with the instruction version number to obtain the version consistency determination result; Based on the comparison between the executable direction in the preparation confirmation information and the candidate sub-instruction direction of the corresponding energy storage system in the candidate sub-instruction set, the execution direction consistency determination result is obtained, and the lower limit of the executable power range and the upper limit of the executable power range are extracted from the preparation confirmation information as executable capability input; The admission weight is calculated based on the current value of the adaptive participation factor that matches the energy storage identifier in the adaptive participation factor table and the reliability parameter. The admission weight is then compared with the admission threshold to obtain the admission determination result. When the preparation confirmation status is confirmed, the version consistency determination result is consistent, the execution direction consistency determination result is consistent, and the admission determination result is passed, the corresponding energy storage identifier is included in the consistency set. If any determination result does not meet the inclusion criteria or the corresponding energy storage system fails to return preparation confirmation information within the preparation timeout threshold, the corresponding energy storage identifier will be included in the frozen set. Issue a commit command to the consistent set and an undo command to the frozen set.

[0011] In one possible implementation, the step involves calculating the grid connection point power deviation based on the plant-level frequency regulation target and the real-time acquired grid connection point power value; using the adaptive participation factor table as an allocation constraint, allocating the grid connection point power deviation to the energy storage systems that have issued submission instructions, and generating a set of correction sub-instructions corresponding to the instruction version number, including: Based on the plant-level frequency regulation target and the real-time acquired grid connection point power value, the grid connection point power deviation is calculated, and the deviation direction of the grid connection point power deviation is determined. Read the set of submitted sub-instructions according to the instruction version number, extract the instruction value and instruction direction of each energy storage system in the consistency set, and combine the lower limit and upper limit of the executable power range of each energy storage system recorded in the preparation confirmation information set to calculate the remaining upward adjustment and remaining downward adjustment of each energy storage system relative to the current instruction value. Read the current value of the adaptive participation factor of each energy storage system in the consensus set from the adaptive participation factor table, generate the allocation weight of each energy storage system based on the current value of the adaptive participation factor, and normalize the allocation weight. Using the power deviation at the grid connection point as the total correction amount, the total correction amount is initially allocated according to the normalized allocation weight to obtain the initial correction share of each energy storage system. The initial correction fraction is combined with the deviation direction to obtain the initial correction command value for each energy storage system; The initial correction command value is compared with the remaining upward or downward adjustment range of the corresponding energy storage system at the boundary. If the initial correction instruction value exceeds the executable boundary of the corresponding energy storage system, the excess portion will be defined as the correction amount to be reallocated, and the corresponding energy storage system that touches the executable boundary will be removed from the allocation objects in subsequent dynamic adjustments. If the preset redistribution number threshold is not reached, the redistribution correction amount is recalculated and allocated to the energy storage systems that have not been removed according to the adaptive participation factor table, until the redistribution correction amount is zero, or all energy storage systems in the consensus set reach the executable power range boundary. The energy storage identifier, instruction version number, and final correction instruction value of each energy storage system in the consensus set are encapsulated to generate a correction sub-instruction set, and the correction sub-instruction set is made to have the same instruction version number as the submitted sub-instruction set.

[0012] In one possible implementation, the step of calculating and generating an instruction execution effectiveness evaluation metric based on the set of corrected sub-instructions, the version execution record of the current instruction, the preparation confirmation information set, and the change in grid connection point power deviation includes: A set of correction sub-instructions is sent to the consensus set, the measured power of the grid connection point after the correction sub-instruction set is sent is obtained, and the change in the power deviation of the grid connection point is calculated based on the power deviation of the grid connection point before and after the correction sub-instruction set is sent. Based on the prepared confirmation information set, version execution record, adaptive participation factor table, grid connection point power deviation and the change in grid connection point power deviation, a unit input feature matrix is ​​generated for each energy storage system. The unit input feature matrix includes version consistency determination result, execution direction consistency determination result, lower limit of executable power range, upper limit of executable power range, instruction value in the submitted sub-instruction set, instruction value in the corrected sub-instruction set, set identifier in the version execution record, grid connection point power deviation, change in grid connection point power deviation, and current value of adaptive participation factor; The unit input feature matrix is ​​input into a pre-trained effectiveness evaluation model, which includes two fully connected layers and one output layer. The effectiveness evaluation value corresponding to each energy storage system is obtained by outputting the effectiveness evaluation model.

[0013] In one possible implementation, the step of dynamically updating the adaptive participation factor table based on the effectiveness evaluation metric to obtain an updated adaptive participation factor table includes: The set of submission sub-instructions and the set of correction sub-instructions are merged to form the final set of frequency modulation sub-instructions issued. Based on the effectiveness evaluation quantity corresponding to each energy storage system, the current value of the adaptive participation factor is updated within the allowable value boundary to obtain the updated adaptive participation factor table. Output the updated adaptive participation factor table and the final set of frequency modulation sub-instructions.

[0014] According to a second aspect of this application, a control device for multi-energy storage asynchronous cooperative frequency modulation output is provided, comprising: The first generation module is used to generate plant-level frequency regulation targets based on the upper-level frequency regulation instructions and grid connection point power values ​​corresponding to the multiple energy storage systems connected in parallel at the thermal power plant. The power allocation module is used to determine the instruction version number based on the plant-level frequency regulation target and the preset adaptive participation factor table corresponding to each energy storage system, and to allocate power to the plant-level frequency regulation target within the executable power range of each energy storage system to obtain a set of candidate sub-instructions that are mapped one-to-one with the instruction version number; wherein, the adaptive participation factor table is used to characterize the execution reliability of each energy storage system version. The collection distribution module is used to distribute the instruction version number and the corresponding candidate sub-instruction set to each energy storage system to form a preparation request; The second generation module is used to obtain the set of preparation confirmation information returned by each energy storage system in response to the preparation request, and to perform a consistency judgment on the preparation request of each energy storage system based on the set of preparation confirmation information, and to generate a submit command or a cancel command for the corresponding energy storage system according to the judgment result. The instruction issuing module is used to issue submission instructions to the energy storage system that generates the submission command, and to issue cancellation instructions to the energy storage system that generates the cancellation command. The power calculation module is used to calculate the grid connection point power deviation based on the plant-level frequency regulation target and the real-time acquired grid connection point power value; and to allocate the grid connection point power deviation to the energy storage system that has issued the submission instruction, using the adaptive participation factor table as the allocation constraint, and to generate a set of correction sub-instructions corresponding to the instruction version number. The instruction correction module is used to issue the correction sub-instruction set to the energy storage systems within the consensus set; The third generation module is used to calculate the evaluation value of the execution effectiveness of the generated instruction based on the set of corrected sub-instructions, the version execution record of the current instruction, the set of preparation confirmation information, and the change in the power deviation of the grid connection point. The dynamic update module is used to dynamically update the adaptive participation factor table according to the effectiveness evaluation quantity, obtain the updated adaptive participation factor table, and apply the updated adaptive participation factor table to the instruction generation and power allocation of the next plant-level frequency regulation target.

[0015] According to a third aspect of this application, an electronic device is provided, comprising: At least one processor; and a memory communicatively connected to the at least one processor; wherein, The memory stores instructions that can be executed by the at least one processor to enable the at least one processor to perform the method described in this application.

[0016] According to a fourth aspect of this application, a non-transitory computer-readable storage medium is provided storing computer instructions for causing the computer to perform the methods described in this application.

[0017] According to a fifth aspect of this application, a computer program product is provided, comprising a computer program or instructions that, when executed by a processor, implement the method described in this application.

[0018] By utilizing the technical solution of this application, the grid connection point deviation correction is performed in a closed loop only within this set, thereby avoiding grid connection point frequency modulation output distortion caused by asynchronous reception, inconsistent directions, or parallel new and old instructions, and improving the consistency and traceability of grid connection point frequency modulation execution.

[0019] It should be understood that the description in this section is not intended to identify key or essential features of the embodiments of this application, nor is it intended to limit the scope of this application. Other features of this application will become readily apparent from the following description. Attached Figure Description

[0020] The above and other objects, features, and advantages of exemplary embodiments of this application will become readily apparent from the following detailed description taken in conjunction with the accompanying drawings. Several embodiments of this application are illustrated in the drawings by way of example and not limitation, in which: In the accompanying drawings, the same or corresponding reference numerals indicate the same or corresponding parts.

[0021] Figure 1 This illustration shows the implementation flow of the control method for asynchronous frequency modulation output of multiple energy storage systems in the embodiments of this application. Figure 1 ; Figure 2 This illustration shows the implementation flow of the control method for asynchronous frequency modulation output of multiple energy storage systems in the embodiments of this application. Figure 2 ; Figure 3 This illustration shows the implementation flow of the control method for asynchronous frequency modulation output of multiple energy storage systems in the embodiments of this application. Figure 3 ; Figure 4 This illustration shows the implementation flow of the control method for asynchronous frequency modulation output of multiple energy storage systems in the embodiments of this application. Figure 4 ; Figure 5 This paper illustrates a schematic diagram of the consensus set and the submission command in an embodiment of this application. Figure 6 This illustration shows the implementation flow of the control method for asynchronous frequency modulation output of multiple energy storage systems in the embodiments of this application. Figure 5 ; Figure 7 This illustration shows a schematic diagram of the alignment of the submission command in an embodiment of this application; Figure 8 This illustration shows the implementation flow of the control method for asynchronous frequency modulation output of multiple energy storage systems in the embodiments of this application. Figure 6 ; Figure 9 This illustration shows a schematic diagram of allocating and generating a set of correction sub-instructions only within the consistency set in an embodiment of this application; Figure 10 This illustration shows the implementation flow of the control method for asynchronous frequency modulation output of multiple energy storage systems in the embodiments of this application. Figure 7 ; Figure 11 A schematic diagram of the version execution record in an embodiment of this application is shown; Figure 12The diagram illustrates the implementation block diagram of the control method for asynchronous frequency modulation output of multiple energy storage systems in an embodiment of this application. Figure 13 A schematic diagram of the composition structure of the electronic device in an embodiment of this application is shown. Detailed Implementation

[0022] To make the objectives, features, and advantages of this application more apparent and understandable, the technical solutions in the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0023] In the following description, the terms "first" and "second" are used merely to distinguish similar objects and do not represent a specific ordering of objects. It is understood that "first" and "second" may be interchanged in a specific order or sequence where permitted, so that the embodiments of this application described herein can be implemented in an order other than that illustrated or described herein.

[0024] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application belongs. The terminology used herein is for the purpose of describing embodiments of this application only and is not intended to limit this application.

[0025] The following description, in conjunction with the accompanying drawings, introduces a control method, device, and electronic equipment for multi-energy storage asynchronous coordinated frequency modulation output provided in this application.

[0026] like Figure 1 As shown, this application provides a control method for asynchronous frequency modulation output of multiple energy storage systems, the method comprising: S101, Based on the obtained upper-level frequency regulation commands and grid connection point power values ​​corresponding to multiple energy storage systems connected in parallel at the thermal power plant, generate plant-level frequency regulation targets; S102, based on the plant-level frequency regulation target and the preset adaptive participation factor table corresponding to each energy storage system, determine the instruction version number, and within the executable power range of each energy storage system, allocate power to the plant-level frequency regulation target to obtain a set of candidate sub-instructions that are mapped one-to-one with the instruction version number; wherein, the adaptive participation factor table is used to characterize the execution reliability of each energy storage system version; the adaptive participation factor table is a pre-initialized weight table that is iteratively updated with the execution effect of the instruction; S103, The instruction version number and the corresponding set of candidate sub-instructions are sent to each energy storage system to form a preparation request; S104, obtain the preparation confirmation information set returned by each energy storage system in response to the preparation request, and perform consistency judgment on the preparation request of each energy storage system based on the preparation confirmation information set, and generate a submit command or cancel command for the corresponding energy storage system according to the judgment result. S105, issue a submission command to the energy storage system corresponding to the submission command and issue a cancellation command to the energy storage system that generated the cancellation command; S106, calculate the grid connection point power deviation based on the plant-level frequency regulation target and the real-time acquired grid connection point power value; use the adaptive participation factor table as the allocation constraint to allocate the grid connection point power deviation to the energy storage system that has issued the submission instruction, and generate a set of correction sub-instructions corresponding to the instruction version number; S107, issue the set of correction sub-instructions to the energy storage systems within the consensus set; S108, calculate and generate an instruction execution effectiveness evaluation quantity based on the set of corrected sub-instructions, the version execution record of this instruction, the set of preparation confirmation information, and the change in the power deviation of the grid connection point; S109, dynamically update the adaptive participation factor table according to the effectiveness evaluation quantity to obtain the updated adaptive participation factor table, and apply the updated adaptive participation factor table to the instruction generation and power allocation of the next plant-level frequency regulation target.

[0027] The technical solution provided by this invention is an improved method for consistent execution of multi-energy storage coordinated frequency regulation commands. By introducing command version numbers and a two-stage submission and orchestration mechanism at the plant-level coordinator side, consistent command effectiveness within the same time period is taken as a prerequisite for forming coordinated frequency regulation power. Candidate sub-commands are not directly triggered for execution. Instead, a preparation request carrying the command version number is first issued. Based on the preparation confirmation information returned by the energy storage system, version consistency and execution direction consistency are determined for each energy storage system. The admission weight is calculated by combining adaptive participation factors, reliability parameters, preparation response delay, and margin of candidate commands within the executable power range. Energy storage units that meet the admission threshold are included in the consistency set, and the remaining units are included in the freeze set. Compared with the traditional method of directly broadcasting commands according to capacity or fixed participation factors, this invention can eliminate abnormal units with asynchronous reception, cross-version parallelism, and opposite execution directions at the execution entry point. This makes the boundaries of units participating in coordinated output under the same version computable and verifiable, effectively reducing the output deviation between the net output at the grid connection point and the plant-level frequency regulation target caused by asynchronous execution.

[0028] This application proposes a version execution record mechanism bound to the submit / revoke execution mechanism. It uses a consensus set as the sole object of the submit instruction and a frozen set as the explicit object of the revoke instruction. Execution window parameters are configured in the submit command to align the instruction version number with the effective period. After the instruction is submitted, measured power information of the grid-connected point is collected and a version execution record is formed. This record at least associates the instruction version number, the boundary between the consensus set and the frozen set, the submitted sub-instruction set, the receipt status, and the grid-connected point power data within the execution window, establishing a traceable correspondence between the grid-connected point observations and the actual set of units participating in the execution. Compared to methods that rely solely on unit receipts or perform closed-loop tracking only at the individual energy storage side, this invention provides a stable and reliable data boundary for subsequent power deviation calculations and execution effect attribution, reducing the problem of lag units overlapping or canceling out new instructions under the influence of old instructions. This eliminates the engineering assumption that the frequency regulation control closed loop relies on the natural synchronization of each controller.

[0029] In addition, this application employs a consistent set-based version error allocation and adaptive participation factor closed-loop update method. After calculating the grid connection point power deviation, it iteratively allocates the deviation amount only within the consistent set, under the constraints of the executable power range and the adaptive participation factor, generating a set of corrected sub-instructions with the same version number as the original submitted instruction. This avoids allocating the corrected amount to frozen units, which could lead to the parallel execution of old version instructions. Furthermore, it constructs a unified feature input using preparation confirmation information, version execution records, and grid connection point power deviation changes. Through the effectiveness evaluation model, it outputs the instruction execution effectiveness evaluation quantity for each energy storage system, thereby driving the adaptive participation factor to dynamically update within the allowed value boundaries. This ensures that subsequent admission weight calculation and deviation allocation weights can truly reflect the actual effectiveness of each unit under version switching and asynchronous execution conditions. This invention integrates consistent execution boundary determination, version-based closed-loop correction, and reliability adaptive update into a complete control link, significantly improving the executability of plant-level frequency regulation instructions and the consistency of grid connection point output for multiple energy storage systems in asynchronous scenarios.

[0030] In some embodiments, such as Figure 2 As shown, the generation of plant-level frequency regulation targets based on the acquired upper-level frequency regulation commands and grid connection point power values ​​corresponding to multiple parallel energy storage systems in a thermal power plant includes: S201, Receive the frequency modulation command from the upper level, extract the command arrival time and the command power target value, and use the command arrival time as the alignment reference for this frequency modulation cycle; S202, collect grid connection point power values ​​within this frequency modulation cycle, perform time-series alignment of the grid connection point power values ​​according to the alignment reference, and use a preset sampling window to perform a sliding average of the grid connection point power values ​​to obtain the grid connection point power reference value; S203, Generate a plant-level frequency regulation target based on the commanded power target value and the grid connection point power reference value; wherein, the plant-level frequency regulation target includes: execution direction and amplitude limit; the amplitude limit is constrained by a preset plant-level adjustable power boundary.

[0031] As an example, this application first receives a frequency regulation command from the grid dispatch center, extracts the command arrival time and the command power target value from the command message, and uses the command arrival time as the timing alignment benchmark for this frequency regulation cycle to unify the command execution timing of each energy storage system. Then, within this frequency regulation cycle, the measured power value of the grid-connected point of the thermal power plant is collected in real time at a fixed sampling frequency. The collected grid-connected point power value is time-aligned according to the command arrival time to eliminate timing deviations caused by communication delays and sampling asynchrony. Then, a sampling window of preset length is used to perform moving average filtering on the time-aligned grid-connected point power value to remove high-frequency noise and instantaneous fluctuations, obtaining a stable and reliable grid-connected point power benchmark value. Finally, based on the command power target value of the upper-level frequency regulation command and the grid-connected point power benchmark value, a plant-level frequency regulation target is calculated and generated. The plant-level frequency regulation target includes an execution direction and a fixed amplitude limit. The amplitude limit is constrained by preset upper and lower boundaries of the plant-level adjustable power to ensure that the plant-level frequency regulation target is within the safe and adjustable range of the system.

[0032] For example, the plant-level coordinator is deployed within the plant-level control network of the thermal power plant. The plant-level coordinator is connected to the upper-level dispatching side through a frequency modulation communication link. The plant-level coordinator is connected to the energy storage controllers corresponding to multiple sets of energy storage through an industrial Ethernet network within the station. The grid connection point power acquisition device periodically sends the measured power data of the grid connection point with timestamps to the plant-level coordinator.

[0033] The plant-level coordinator receives frequency modulation commands from the upper-level unit and extracts the command arrival time from the commands. Compared with the commanded power target value and construct instruction data set The plant-level coordinator will Write the status of this frequency modulation cycle as the alignment reference for this frequency modulation cycle; at the same time, set the duration of this frequency modulation cycle to a preset value. and time interval The effective acquisition interval for this frequency regulation cycle is determined to constrain the attribution of the sampled data of the measured power at the grid connection point, and to avoid cross-cycle data from being mixed in and affecting the consistency of the instruction version number boundary.

[0034] Plant-level coordinator in time interval Internally collected measured power at grid connection points; each sample includes the sampling time. With power value ,in This represents the measured power at the grid connection point. The plant-level coordinator organizes the sampling sequence into a sampling matrix of the measured power at the grid connection point. Each row of the matrix corresponds to one sampled data point, and the first column is... The second column is The plant-level coordinator uses a preset sampling period. Generate aligned time vector The vector elements are ,in The integers are non-negative and the alignment times fall within the time interval. Plant-level coordinator based on right Perform timing alignment to obtain the aligned power vector. The vector elements are the power values ​​corresponding to each alignment time; when a certain alignment time is not aligned with... When the sampling times coincide, the plant-level coordinator selects two sampled data points before and after the alignment time, performs linear interpolation, generates the power value at that alignment time, and writes it into the system. The plant-level coordinator uses a preset number of sampling window points. right Perform a moving average, which is calculated by averaging the most recent... The arithmetic mean of the aligned power values ​​is used to obtain the reference value of the measured power at the grid connection point. .

[0035] Plant-level coordinator according to and Generate plant-level frequency modulation target ,in This represents the target adjustment amount for the plant-level frequency regulation objective. The plant-level coordinator uses... The positive or negative sign determines the command direction, and the command direction is recorded as a direction identifier. ,in It selects one of three states: upward adjustment, downward adjustment, or zero direction. The plant-level coordinator... Amplitude limiting is applied, and the limits are defined by preset plant-level adjustable power limits, with the upper limit being... The lower boundary is ;when Exceed When Set as ,when Below When Set as and keep Compared to the limited amplitude Consistency is used to ensure that the direction of the candidate sub-instructions in the candidate sub-instruction set has the same judgment criteria as the executable direction in the preparation confirmation information.

[0036] The plant-level coordinator reads the adaptive participation factor table from local storage. The adaptive participation factor table records the current value, allowable value boundaries, and reliability parameters of each adaptive participation factor, categorized by energy storage identifier. The energy storage identification of the set of energy storage is as follows: The current value of the adaptive participation factor is denoted as The allowed value boundaries are denoted as The reliability parameter is denoted as The plant-level coordinator transforms the adaptive participation factor table into a computationally oriented structured data set, including energy storage identification vectors. Adaptive participation factor vector Allowable value boundary matrix With reliability parameter vector ,in For energy storage capacity, The Line contains and Plant-level coordinator maintenance , , and The index consistency ensures that the energy storage identifier and the participation factor entry are matched one-to-one in the calculation of admission weights and the allocation of grid connection point power deviation.

[0037] In some embodiments, such as Figure 3 As shown, based on the plant-level frequency regulation target and a preset adaptive participation factor table corresponding to each energy storage system, the instruction version number is determined. Within the executable power range of each energy storage system, power allocation is performed on the plant-level frequency regulation target to obtain a set of candidate sub-instructions that are mapped one-to-one with the instruction version number, including: S301, Read the pre-stored adaptive participation factor table in the plant-level coordinator to obtain the corresponding entries for each energy storage system; the entries include the energy storage identifier, the current value of the adaptive participation factor, the allowed value boundary, and the reliability parameter used to characterize the reliability of version execution; S302, determine the instruction boundary of the current frequency modulation cycle according to the plant-level frequency modulation target, and combine the instruction boundary with the execution direction of the plant-level frequency modulation target and the amplitude limit value to generate an instruction version number; S303, based on the current value of the adaptive participation factor and the plant-level frequency regulation target, perform initial power allocation to obtain the initial decomposition command value corresponding to each energy storage system; S304, based on the executable power range corresponding to each energy storage system, the initial decomposition command value is limited, and the remaining power generated by the limiting is redistributed to the energy storage system that has not touched the boundary of the executable power range according to the adaptive participation factor table until the decomposition command of all energy storage systems meets the executable power range constraint. S305 encapsulates the energy storage identifier, instruction version number, instruction direction, and decomposed instruction value of each energy storage system into a set of candidate sub-instructions.

[0038] As an example, firstly, the adaptive participation factor table pre-stored in the plant-level coordinator is read to obtain the configuration entry corresponding to each energy storage system. This entry includes at least the energy storage identifier, the current value of the adaptive participation factor, the allowed value boundary of the adaptive participation factor, and a reliability parameter characterizing the execution reliability of each energy storage system version. Then, the instruction boundary for the current frequency regulation cycle is determined based on the plant-level frequency regulation target. This instruction boundary is combined with the execution direction of the plant-level frequency regulation target and the amplitude limit value to generate a unique corresponding instruction version number, used to distinguish frequency regulation instructions for different cycles and different targets. Next, initial power allocation is performed based on the current value of the adaptive participation factor for each energy storage system and the plant-level frequency regulation target to obtain the initial decomposition instruction value corresponding to each energy storage system. Then, the initial decomposition instruction value is constrained according to the current executable power range of each energy storage system. The remaining power generated after the constraint is redistributed to energy storage systems that have not reached the executable power range boundary according to the adaptive participation factor table. This iterative allocation continues until the decomposition instructions of all energy storage systems satisfy the executable power range constraint. Finally, the energy storage identifier, instruction version number, instruction direction, and the final decomposed instruction value of each energy storage system are encapsulated to form a set of candidate sub-instructions that are mapped one-to-one with the instruction version number.

[0039] For example, the plant-level coordinator uses the plant-level frequency modulation target. Directional signs Command arrival time Frequency modulation cycle duration Energy storage identification vector With adaptive participation factor vector For input, generate instruction version number With the set of candidate sub-instructions.

[0040] The plant-level coordinator determines the command boundary for this frequency modulation cycle, with the command boundary starting at [time]. The end time of the instruction boundary is denoted as , By and The sum is obtained; the instruction boundaries are organized into instruction boundary vectors. Plant-level coordinator maintenance version number The version number is an incrementing integer. When a new frequency modulation command is received from the superior and completed... During the update Add one. The plant-level coordinator will , and The version field vector is composed according to a fixed field order. ;Will and Convert to a millisecond timestamp integer. Mapped to direction code The upward adjustment direction corresponds to The downward adjustment direction corresponds to Zero direction corresponds ;Will , millisecond integers, Millisecond integers and The instruction version number is obtained by serializing the data into a fixed-byte length and concatenating them in order. Command version number It serves as a common index field for preparation requests, commit commands, undo commands, version execution records, and sets of correction sub-instructions.

[0041] Plant-level coordinator To decompose the input, calculate the sum of adaptive participation factors. , for The summation results of each element; the initial decomposition command value is calculated for each energy storage unit. , By Multiply and divide by The plant-level coordinator organizes the initial decomposition instruction values ​​into an initial decomposition instruction vector. And the direction of the candidate sub-instructions is uniformly set to This allows for a direct comparison of the candidate sub-instruction direction with the executable direction to ensure consistent execution direction.

[0042] The plant-level coordinator obtains the executable power range for each energy storage unit. The plant-level coordinator sends a capacity status request to each energy storage controller; the capacity status request includes the energy storage identifier. With the request time ,in Pick The subsequent preset offset time; the energy storage controller returns a capability status response, which includes the lower limit of the executable power range. With the upper limit of the executable power range The plant-level coordinator organizes the executable power ranges of each energy storage system into an executable power range matrix. , The Line contains .

[0043] Plant-level coordinator Perform limiting based on the executable power range to obtain the limited instruction vector. ,in By Cut to The range is obtained. The plant-level coordinator calculates the residual power generated by the limiting. , By minus The summation of each element yields the result. The plant-level coordinator constructs a reassignable identifier vector. , among which when To adjust the direction and Time to take ,when To adjust the direction and Time to take In other cases, take The plant-level coordinator calculates the remaining adjustable amplitude vector. , among which when When adjusting the direction upwards Depend on Get, when When adjusting the direction downwards Depend on get.

[0044] Plant-level coordinator's maximum number of reassignment attempts Perform remaining power reallocation under constraints, with a maximum number of reallocation attempts. This is a preset integer. During each reallocation, the plant-level coordinator uses... Select energy storage items to participate in the reallocation and calculate the sum of participation factors. , To meet of Summation result; for each satisfying Energy storage calculation allocation increment , By according to exist The proportion in is allocated to obtain the result; and Perform boundary comparisons when When Set as ,when When Set as ;Will Overlay After the update And according to the updated refresh and Then press the updated version Recalculate .when Zero, or All elements are zero, or reach Stop reallocation when the time comes, and obtain the decomposed instruction value vector. , From the time of stopping Sure.

[0045] Plant-level coordinator based on Command version number Command direction and Construct a set of candidate subinstructions. For the first... Energy storage generates candidate sub-instruction line vectors ,in for The Each element; All candidate sub-instruction row vectors are arranged into a candidate sub-instruction set according to the energy storage index order. The candidate sub-instruction set is used for consistency orchestration during the preparation phase to generate instructions carrying... Prepare the request and conduct version consistency determination and execution direction consistency determination.

[0046] In some embodiments, the formation preparation request includes: Write the energy storage identifier, instruction version number, corresponding candidate sub-instruction direction, and corresponding candidate sub-instruction value into the preparation request, and configure a preparation timeout threshold for the preparation request. The preparation request is sent to each energy storage system, and a timer based on the preparation timeout threshold is started on the plant-level coordinator side.

[0047] As an example, a preparation request data frame is constructed, and the core information corresponding to each energy storage system is written into the preparation request one by one. Specifically, this includes: energy storage identifier (used to uniquely distinguish each parallel energy storage system, such as "ESS-01", "ESS-02", etc.), instruction version number of the current frequency regulation cycle (consistent with the version number corresponding to the candidate sub-instruction set to ensure instruction traceability), candidate sub-instruction direction corresponding to the energy storage system (i.e., frequency regulation execution direction, such as charging or discharging), and candidate sub-instruction value corresponding to the energy storage system (i.e., the decomposed frequency regulation power instruction value). At the same time, a preparation timeout threshold is configured for the preparation request. Considering the communication latency of multiple energy storage systems in the thermal power plant and the response speed of the energy storage controller, the preset timeout threshold is 500ms. If the energy storage system does not return preparation confirmation information within this threshold, it is determined that the preparation has failed, so as to avoid the overall frequency regulation timing being affected by the response lag of individual energy storage systems. The plant-level coordinator distributes the completed preparation requests to each energy storage system via a pre-defined communication link (such as industrial Ethernet or a dedicated 4G / 5G communication module). This ensures that each energy storage system accurately receives its corresponding preparation request, preventing any mis-sending or missed commands. Simultaneously, a timing module is activated on the plant-level coordinator side. Starting from the moment the preparation request is sent, the timing begins based on a pre-defined preparation timeout threshold, monitoring the real-time return of preparation confirmation information from each energy storage system. This provides a timing basis for subsequent determination of whether an energy storage system should be included in the consensus set or the frozen set.

[0048] This application does not directly trigger execution of the candidate sub-instruction set. Instead, it first generates a preparation request carrying the instruction version number. The preparation request drives each energy storage unit to output preparation confirmation information and summarizes the reception status of each energy storage unit for the instruction version number, the current executable direction, and the executable power range to form a preparation confirmation information set. Based on the preparation confirmation information set, each energy storage unit first performs a version consistency determination to identify whether the instruction version number in the preparation confirmation information matches the instruction version number. Then, each energy storage unit performs an execution direction consistency determination to identify whether the executable direction in the preparation confirmation information matches the corresponding sub-instruction direction in the candidate sub-instruction set. The version consistency determination and execution direction consistency determination are used to constrain two types of distortion sources: parallel execution of different instruction version numbers and execution in opposite directions within the same version.

[0049] In some embodiments, such as Figure 4 As shown, the step of obtaining the preparation confirmation information set returned by each energy storage system in response to the preparation request, and performing a consistency judgment on the preparation request of each energy storage system based on the preparation confirmation information set, and generating a submit command or cancel command for the corresponding energy storage system according to the judgment result, includes: S401: Within the preparation timeout threshold, receive preparation confirmation information returned by each energy storage system and summarize it into a preparation confirmation information set according to the energy storage identifier. S402, extract the receiving instruction version number and preparation confirmation status of each energy storage system from the preparation confirmation information set, compare the receiving instruction version number with the instruction version number, and obtain the version consistency determination result. S403, based on the comparison between the executable direction in the preparation confirmation information and the candidate sub-instruction direction of the corresponding energy storage system in the candidate sub-instruction set, a consistent execution direction determination result is obtained, and the lower limit of the executable power range and the upper limit of the executable power range are extracted from the preparation confirmation information as executable capability input; S404, calculate the admission weight based on the current value of the adaptive participation factor that matches the energy storage identifier in the adaptive participation factor table and the reliability parameter, and compare the admission weight with the admission threshold to obtain the admission determination result; S405, in response to the preparation confirmation status being confirmed, the version consistency determination result being consistent, the execution direction consistency determination result being consistent, and the admission determination result being passed, the corresponding energy storage identifier is included in the consistency set. S406, in response to any determination result not meeting the inclusion condition or the corresponding energy storage system not returning preparation confirmation information within the preparation timeout threshold, the corresponding energy storage identifier is included in the frozen set. S407 issues a commit command to the consistent set and a revocation command to the frozen set.

[0050] As an example, the plant-level coordinator continuously monitors the return of preparation confirmation information from each energy storage system. It only accepts confirmation information returned within the preparation timeout threshold. Energy storage systems that fail to return confirmation information within the timeout period are directly marked as preparation failures. For validly returned preparation confirmation information, it categorizes and summarizes them according to their energy storage identifiers to form a complete preparation confirmation information set. This ensures that the confirmation information of each energy storage system accurately corresponds to its own identifier, facilitating subsequent individual assessments. From the preparation confirmation information set, the receiving command version number and preparation confirmation status of each energy storage system are extracted. The extracted receiving command version number is precisely compared with the command version number of the current frequency regulation cycle. If the two are completely identical, the version consistency determination result is "consistent"; if there are version number mismatches, missing numbers, or other issues, the determination result is "inconsistent". Extract the executable direction from each energy storage system in the preparation confirmation information and compare it with the candidate sub-instruction direction of the corresponding energy storage system in the candidate sub-instruction set. If both are charging or discharging, the execution direction is consistent, and the result is "consistent"; if the directions are opposite, it is "inconsistent". At the same time, extract the lower limit and upper limit of the current executable power range of each energy storage system from the preparation confirmation information as the executable capability input data for subsequent power allocation and verification. Read the adaptive participation factor table pre-stored by the plant-level coordinator, match the current value of the adaptive participation factor and reliability parameter (e.g., 0.95) corresponding to each energy storage system according to the energy storage identifier, and calculate the admission weight of each energy storage system using a preset algorithm (e.g., weighted summation algorithm). The preset admission threshold is 0.6. Compare the calculated admission weight with the admission threshold. If the admission weight is ≥0.6, the admission result is "passed"; if <0.6, it is "failed". The system comprehensively evaluates four criteria for each energy storage system: when the preparation confirmation status is "confirmed," the version consistency judgment result is "consistent," the execution direction consistency judgment result is "consistent," and the access judgment result is "passed," the energy storage system's identifier is included in the consistency set, indicating it possesses the capability for coordinated frequency regulation execution. If any of the above conditions are not met, or if the energy storage system fails to return preparation confirmation information within the 500ms preparation timeout threshold, its energy storage identifier is included in the freeze set, indicating it lacks the capability for coordinated frequency regulation execution. The plant-level coordinator sends a submission command to all energy storage systems in the consistency set via a pre-defined communication link. This command includes core information such as the instruction version number and the final execution instruction, instructing them to prepare for execution according to the candidate sub-instructions. Simultaneously, it sends a cancellation command to all energy storage systems in the freeze set, instructing them to abandon the preparation work for this candidate sub-instruction, maintain their original operating status, and avoid invalid instructions affecting the overall frequency regulation effect.

[0051] In this application, after completing two types of consistency determinations, the admission weights are further calculated based on the adaptive participation factor table, and the admission weights are combined with the admission thresholds to form admission conditions. The admission weights are used to express the reliability of each energy storage unit's effective contribution under the plant-level frequency regulation cycle, and the reliability is associated with the real-time executable status of the preparation confirmation information set, so that the determination of the consistency set is no longer equivalent to inclusion in execution upon receiving confirmation. When the admission weights meet the admission thresholds and the preparation confirmation information meets the version consistency determination and execution direction consistency determination, a consistency set is output; when either condition is not met, a frozen set is output. The frozen set is used to separate energy storage units with uncertain execution from the resultant power path of this instruction version number, establishing a traceable relationship between the measured power at the grid connection point and the net output formed by the consistency set, providing clear object boundaries for submitting and revoking commands. The preparation confirmation information set, the consistency set, and the frozen set together constitute the verifiable output of the consistency orchestration in the preparation phase, ensuring that the subsequent submission phase does not rely on controller synchronization assumptions.

[0052] For example, the plant-level coordinator uses the instruction version number Candidate subinstruction set Energy storage identification vector Adaptive participation factor vector With reliability parameter vector As input, perform a preparation phase consistency orchestration on the candidate subinstruction set, where For the amount of energy stored. Set of candidate sub-instructions. Depend on The candidate sub-instruction row vectors are composed of the first row vector. The candidate sub-instruction row vector is denoted as ,in Indicates the direction of the candidate sub-instruction. Indicates the candidate sub-instruction value.

[0053] Plant-level coordinator according to and Generate a request matrix , The Travelogue ,in This indicates the timeout threshold for preparation. The plant-level coordinator will... The energy storage identifier is sent line by line to the corresponding energy storage controller, and the ready-to-send vector is recorded. ,in To label energy storage Issued The sending time is used to determine whether the preparation for acknowledgment has timed out. For example... Figure 5As shown, starting from the candidate sub-instruction set and multiple energy storage controllers, two sets with clear boundaries are formed through "consistent orchestration in the preparation phase": the consistent set and the frozen set. In the submission phase, a submission command is issued to the consistent set, enabling it to enter the execution of this version; a revocation command is issued to the frozen set, clarifying that it will not participate in this version. The key point of the diagram is "fixing the set boundaries before execution," rather than all devices running in parallel. This application removes uncertain units such as asynchronous operations, disconnections, and unmet state conditions from the execution path of this version, avoiding the superposition / cancellation of grid connection point power and effect drift caused by parallel execution, thus improving version execution consistency and engineering robustness.

[0054] After receiving the preparation request, the energy storage controller returns a preparation confirmation message, which includes the version number of the received command. 1. Prepare to confirm status Executable direction Lower limit of the executable power range With the upper limit of the executable power range The plant-level coordinator receives preparation confirmation information from all energy storage controllers and summarizes this information according to energy storage identifiers to form a preparation confirmation information matrix. , The Travelogue ,in This indicates that the plant-level coordinator receives the energy storage identifier. Prepare to confirm the time of receiving the information; when the energy storage identifier... In the time interval If no confirmation message is returned, the plant-level coordinator will... Set as unconfirmed, and , , , Set it to an invalid placeholder value, and define the access weight of the energy storage as zero, so that the access determination has a directly calculable numerical input.

[0055] Plant-level coordinator from Extract the received instruction version number vector With the preparation to confirm the state vector ,Will With instruction version number The version consistency determination vector is obtained by comparing each item. ,in It can be either consistent or inconsistent. The plant-level coordinator starts from... Extract the executable direction vector and from Extracting candidate sub-instruction direction vectors ,Will and The execution direction consistency determination vector is obtained by comparing each item. ,in It will take either a consistent or inconsistent state. The plant-level coordinator will... The lower and upper limits of the executable power range are extracted into an executable power range matrix. , The Line contains .

[0056] In this application, during the preparation phase, a consistent set of "version consistent + direction consistent + access passed" is selected, and only instructions of the same version are submitted to the consistent set; the frozen set is removed from isolation and does not participate in the synergy, so that the synergy of the grid connection point is consistent with the target, the output is not distorted and the execution boundary is traceable.

[0057] In some embodiments, such as Figure 6 As shown, the step involves calculating the grid connection point power deviation based on the plant-level frequency regulation target and the real-time acquired grid connection point power value; using the adaptive participation factor table as an allocation constraint, allocating the grid connection point power deviation to the energy storage systems that have issued submission instructions, and generating a set of correction sub-instructions corresponding to the instruction version number, including: S601, based on the plant-level frequency regulation target and the real-time acquired grid connection point power value, calculate the grid connection point power deviation and determine the deviation direction of the grid connection point power deviation; S602, read the set of submitted sub-instructions according to the instruction version number, extract the instruction value and instruction direction of each energy storage system in the consistency set, and combine the lower limit and upper limit of the executable power range of each energy storage system recorded in the preparation confirmation information set to calculate the remaining upward adjustment and remaining downward adjustment of each energy storage system relative to the current instruction value. S603, Read the current value of the adaptive participation factor of each energy storage system in the consensus set from the adaptive participation factor table, generate the allocation weight of each energy storage system based on the current value of the adaptive participation factor, and normalize the allocation weight. S604, using the power deviation at the grid connection point as the total correction amount, the total correction amount is initially allocated according to the normalized allocation weight to obtain the initial correction share of each energy storage system. S605, combine the initial correction share with the deviation direction to obtain the initial correction command value for each energy storage system; S606, compare the initial correction command value with the remaining upward or downward adjustment range of the corresponding energy storage system at the boundary. S607, if the initial correction instruction value exceeds the executable boundary of the corresponding energy storage system, the excess part is defined as the correction amount to be reallocated, and the corresponding energy storage system that touches the executable boundary is removed from the allocation objects of subsequent dynamic adjustment. S608, if the preset redistribution number threshold is not reached, the redistribution correction amount is recalculated and allocated to the energy storage systems that have not been removed according to the adaptive participation factor table, until the redistribution correction amount is zero, or all energy storage systems in the consensus set have reached the executable power range boundary. S609, encapsulate the energy storage identifier, instruction version number and final correction instruction value of each energy storage system in the consistency set to generate a correction sub-instruction set, and ensure that the correction sub-instruction set is consistent with the instruction version number of the submitted sub-instruction set.

[0058] As an example, the plant-level coordinator collects the measured power value of the power plant's grid connection point in real time. Combined with the generated plant-level frequency regulation target, the power deviation of the grid connection point is calculated as 150kW using the formula "power deviation of grid connection point = power value of plant-level frequency regulation target command - power value of grid connection point in real time". At the same time, the deviation direction is determined to be "insufficient power, output needs to be increased", that is, each energy storage system needs to be corrected according to the discharge direction. Based on the instruction version number (V20260317-001) of this frequency regulation cycle, the corresponding set of submitted sub-instructions is read from the plant-level coordinator storage unit. The instruction values ​​and instruction directions (all are discharge directions, with instruction values ​​of 120kW, 100kW, and 80kW respectively) of each energy storage system (assuming ESS-01, ESS-02, and ESS-03) within the consistency set are extracted. At the same time, the executable power range of each energy storage system recorded in the previously prepared confirmation information set is retrieved. The executable power range of ESS-01 is 0~150kW, ESS-02 is 0~120kW, and ESS-03 is 0~100kW. Based on this, the remaining upward adjustment range (discharge amplitude increase) and the remaining downward adjustment range (discharge amplitude decrease) of each energy storage system relative to the current instruction value are calculated. From the adaptive participation factor table pre-stored by the plant-level coordinator, the current values ​​of the adaptive participation factors for each energy storage system within the consensus set are read: ESS-01 corresponds to 0.32, ESS-02 corresponds to 0.28, and ESS-03 corresponds to 0.20. Based on these current values, allocation weights for each energy storage system are generated, ensuring consistency with the current values ​​of the adaptive participation factors. Subsequently, these allocation weights are normalized, and the normalized weights are calculated as follows: ESS-01 is 0.32 / (0.32+0.28+0.20)=0.4, and ESS-02 is 0.28. / 0.8=0.35, and ESS-03 is 0.20 / 0.8=0.25, ensuring that the sum of all weights after normalization is 1.4. Using the calculated grid connection point power deviation of 150kW as the total correction amount, the total correction amount is initially allocated according to the normalized allocation weights, and the initial correction share of each energy storage system is calculated: ESS-01 initial correction share = 150×0.4=60kW, ESS-02 initial correction share = 150×0.35=52.5kW, ESS-03 initial correction share = 150×0.25=37.5kW. The initial correction share of each energy storage system is combined with the deviation direction (increased output, increased discharge amplitude) to obtain the initial correction command value of each energy storage system: ESS-01 initial correction command value = original command value 120kW + initial correction share 60kW = 180kW; ESS-02 initial correction command value = 100kW + 52.5kW = 152.5kW; ESS-03 initial correction command value = 80kW + 37.5kW = 117.5kW.

[0059] The initial correction command values ​​of each energy storage system are compared with the remaining upward adjustment range of the corresponding energy storage system (the deviation direction in this case is to increase output, so only the remaining upward adjustment range needs to be compared): ESS-01 initial correction command value 180kW > its remaining upward adjustment range 30kW (corresponding to the executable upper limit of 150kW), exceeding the executable boundary; ESS-02 initial correction command value 152.5kW > its remaining upward adjustment range 20kW (corresponding to the executable upper limit of 120kW), exceeding the executable boundary; ESS-03 initial correction command value 117.5kW > its remaining upward adjustment range 20kW (corresponding to the executable upper limit of 100kW), exceeding the executable boundary.

[0060] For energy storage systems exceeding the executable boundary, the following adjustments are made to the redistribution amount: ESS-01 excess portion = 180 - 150 = 30kW, so ESS-01 is removed from subsequent redistribution targets; ESS-02 excess portion = 152.5 - 120 = 32.5kW, so ESS-02 is removed from subsequent redistribution targets; ESS-03 excess portion = 117.5 - 100 = 17.5kW, so ESS-03 is removed from subsequent redistribution targets; the current redistribution adjustment amount is 30 + 32.5 + 17.5 = 80kW; the preset redistribution threshold is 3 times, and the current redistribution count is 1, which does not reach the threshold, so redistribution continues.

[0061] It should be noted that since all energy storage systems within the consensus set have reached the limits of their executable power range, further reallocation is impossible. Therefore, the reallocation process is halted. The 80kW reallocation correction amount is temporarily not allocated due to the lack of available allocation targets (this can be compensated for later in the next frequency regulation cycle). The energy storage identifier of each energy storage system within the consensus set, the current instruction version number (V20260317-001), and the final correction instruction value (150kW for ESS-01, 120kW for ESS-02, and 100kW for ESS-03) are encapsulated to generate a correction sub-instruction set. This correction sub-instruction set is ensured to maintain the same instruction version number as the submitted sub-instruction set, facilitating instruction traceability and execution effect monitoring.

[0062] For example, the plant-level coordinator adapts the participation factor vector. Reliability parameter vector The return delay of the confirmation information and the margin of the candidate sub-instruction value within the executable power range are used to calculate the admission weight vector. , among which when For confirmation, the first The access weight for energy storage is calculated using the following formula:

[0063] in, Energy storage label Admission weights; This indicates the current value of the adaptive participation factor; Represents reliability parameters; Indicates the time to confirm receipt of information; Indicates the time when the request is ready to be sent; Indicates the timeout threshold for preparation; Indicates the candidate subinstruction value; Indicates the lower limit of the executable power range; Indicates the upper limit of the executable power range; This indicates taking the smaller value; This indicates that the larger value is taken when compared to zero. The plant-level coordinator... When the last term in the expression is set to zero, the denominator should be set to zero; when When unconfirmed, Set to zero.

[0064] Plant-level coordinator Performing normalization yields the normalized admission weight vector. The normalization method is to normalize each Divide by The summation of all elements; when When the sum of all elements is zero, Set all elements to zero. The plant-level coordinator sets the admission threshold. ,in This represents the admission threshold value. and The admission decision vector is obtained by comparing each item. ,in The result is either a pass or a fail. The preparation information table is shown in Table 1.

[0065] Table 1 Preparation Information Table

[0066] Plant-level coordinator based on , , and Generate a consistent set of identifier vectors With frozen set identifier vector .when To confirm and To be consistent and To be consistent and When it passes, Set to 1 and Set to 0; when any condition is not met, Set to 0 and Set to 1. The plant-level coordinator will Energy storage identifiers of 1 form a consistent set, Energy storage identifiers of 1 form a frozen set and are retained. , , Using two types of set identifier vectors as boundary inputs for commit commands, undo commands, and intra-version error allocation, the instruction version number... Asynchronous reception and opposite-direction execution between energy storage controllers from different manufacturers are explicitly stripped away.

[0067] The plant-level coordinator establishes a communication session with each energy storage controller, addressed by energy storage identifier. The plant-level coordinator uses a set of candidate sub-commands. Consistent set identifier vector Frozen set identifier vector Command version number Prepare and confirm the information matrix The system takes the measured power sampling matrix at the grid connection point as input, executes submit and cancel commands, and generates a version execution record. The energy storage identification vector is denoted as... ,in For the amount of energy stored; and index and Consistent.

[0068] Plant-level coordinator based on right Filter and extract those that meet the requirements. Candidate sub-instruction lines with a corresponding element of 1 form a set of submitted sub-instructions. Submit a set of sub-instructions. It consists of multiple commit sub-instruction row vectors, the first one... The row vector of the submission sub-instruction corresponding to the energy storage unit is denoted as: ,in Indicates the direction of the command. This represents the instruction value corresponding to the energy storage identifier within the submitted sub-instruction set. To support the version detail matrix's row-by-row recording of all energy storage indices, the plant-level coordinator synchronously constructs a full vector of submitted instruction values. , among which when When the corresponding element is 0, the corresponding element will be... Set to zero.

[0069] The plant-level coordinator encapsulates the set of commit sub-instructions into a commit command. Submit command Includes command header vector Execution window parameters With the set of sub-commands ,in Indicates the time when the submission command was generated. This indicates the preset execution window duration. The plant-level coordinator is based on the preparation confirmation information matrix. The established start time for constructing the effective date is the time when the preparation request is sent and the preparation confirmation information is received. And calculate using the following formula:

[0070] in, Indicates the start time of the effective period; Indicates the time when the submission command was generated; Indicates the need to confirm the information matrix. China Energy Storage Label The corresponding preparation confirmation information reception time; This indicates that a request for energy storage identification has been prepared. The preparation request is sent at the specified time. Represents the identity vector of a consistent set The One element; Indicates the duration of effective protection; This represents the maximum value operator. The plant-level coordinator, when satisfying... When the entry is empty, Treat it as zero, Retain the time value as directly calculable. The plant-level coordinator will activate the end time. according to Confirm and write This is used to constrain the effective time boundary of the submitted sub-instruction set.

[0071] The plant-level coordinator unicasts the corresponding energy storage identifiers to the consensus set one by one. And record the submission and sending time vector. , of which only The entry with a value of 1 is written with the valid transmission time. After receiving the submit command, the energy storage controller returns a submit-receive receipt, which contains the energy storage identifier. Receive instruction version number With the receiving identifier ,in The system displays either "received" or "not received." The plant-level coordinator sets a timeout threshold for the receipt. The receipt will be placed at Items arriving within the specified timeframe are recorded as valid receipts; when a valid receipt is received... and When there is a discrepancy, the corresponding Set to "Not Received". The plant-level coordinator will summarize all receipts into a submission receipt matrix. , The Travelogue and from Extract the submission and reception state vector For those who are not in The entry that returns a receipt will Set to not received.

[0072] like Figure 7 As shown, under the same command version number, an execution window parameter (effective start time, effective end time) is introduced. Even if the time when each energy storage controller receives the command is different, the submitted command is constrained to take effect uniformly within the same time window. On the timeline, the "receive time" of each controller is dispersed, but the "submit command effective" interval is aligned, thus ensuring the consistency of the same version on the parallel side. This application can suppress the dispersed execution caused by communication delay / asynchronous reception, reduce the distortion caused by the instantaneous superposition or cancellation of power at the grid connection point, and make the frequency regulation response more predictable and more in line with the planned curve.

[0073] The plant-level coordinator encapsulates the energy storage identifier and instruction version number corresponding to the frozen set into a revocation command matrix. , The Travelogue , of which only Generate a valid row for entry 1; the plant-level coordinator will The command is unicast row-by-row to the energy storage controllers corresponding to the frozen sets. Upon receiving the revocation command, the energy storage controller locates the relevant data based on its local command version number cache. The bound pending instructions are cleared, and the process exits. The corresponding execution path is then executed, and a cancellation receipt is returned, which includes the energy storage identifier. Receive instruction version number With the receiving identifier The plant-level coordinator will aggregate the revocation receipts into a revocation receipt matrix. , The Travelogue And form a cancel reception state vector. For those who are not in The entry that returns a receipt will Set to not received.

[0074] Plant-level coordinator at the end of effective period Upon arrival, obtain the measured power sampling matrix at the grid connection point. , Each line contains the sampling time. With power value ,in This indicates the measured power at the grid connection point. The plant-level coordinator operates within the time interval. Internal screening Obtain the sampling matrix within the window The number of samples within the window is denoted as Plant-level coordinator The arithmetic mean of the power columns in the formula is used to obtain the measured power at the grid connection point within the version. and will With instruction version number pass The alignment binding is completed during the effective period.

[0075] The plant-level coordinator generates version execution records, which contain version header information vectors. Version Details Matrix Version Details Matrix Record each line according to its energy storage label. The Travelogue ,in This indicates two states: a set identifier and a consistent set or a frozen set. Taken from The submission and receipt status is taken from The cancellation of the receiving status is taken from This causes the instruction version number to be... The consistent set boundary, frozen set boundary, and committed sub-instruction set are directly referenced in the same record structure as the measured power of the grid connection point within the version.

[0076] In some embodiments, such as Figure 8 As shown, the step of calculating and generating an instruction execution effectiveness evaluation metric based on the set of corrected sub-instructions, the version execution record of this instruction, the set of preparation confirmation information, and the change in grid connection point power deviation includes: S801, send a set of correction sub-instructions to the consensus set, obtain the measured power of the grid connection point after the correction sub-instruction set is sent, and calculate the change in the power deviation of the grid connection point based on the power deviation of the grid connection point before and after the correction sub-instruction set is sent. S802, Based on the prepared confirmation information set, version execution record, adaptive participation factor table, grid connection point power deviation and the change in grid connection point power deviation, generate a unit input feature matrix for each energy storage system corresponding to a sample. S803, the unit input feature matrix includes version consistency determination result, execution direction consistency determination result, lower limit of executable power range, upper limit of executable power range, instruction value in the submitted sub-instruction set, instruction value in the corrected sub-instruction set, the set identifier in the version execution record, grid connection point power deviation, change in grid connection point power deviation, and current value of adaptive participation factor; S804, the unit input feature matrix is ​​input into the pre-trained effectiveness evaluation model, the effectiveness evaluation model includes two fully connected layers and one output layer, and the effectiveness evaluation quantity corresponding to each energy storage system is obtained through the output of the effectiveness evaluation model.

[0077] As an example, the plant-level coordinator, through the redundant design of the industrial Ethernet main link and 4G / 5G backup link described above, synchronously issues a set of correction sub-instructions to ESS-01, ESS-02, and ESS-03 within the consensus set. Immediately after issuance, the issuance time is recorded to ensure synchronous effectiveness. Simultaneously, the measured power of the grid-connected points is continuously collected at a sampling frequency of 10Hz, with 10 sets of data collected consecutively. After processing using the moving average method, the average measured power of the grid-connected points after the issuance of the correction sub-instruction set is obtained. The grid-connected point power deviation before the issuance of the correction sub-instruction set is retrieved and, combined with the measured power of the grid-connected points after issuance, is recalculated. The preparation confirmation information set, version execution record, and adaptive participation factor table for this frequency regulation cycle are retrieved and combined with the calculated grid-connected point power deviation. For each energy storage system within the consensus set, a sample unit input feature matrix is ​​generated to ensure a one-to-one correspondence between the feature matrix and the energy storage system, without confusion. The unit input feature matrices corresponding to the three energy storage systems are input into the pre-trained effectiveness evaluation model in the plant-level coordinator. The effectiveness evaluation quantity is generated using an effectiveness evaluation model, which is deployed on the plant-level coordinator side. This model is used to convert the observable states in the preparation confirmation information set and version execution record into effectiveness evaluation quantities that can be directly used to adjust the adaptive participation factor table. The effectiveness evaluation model works with one sample per energy storage system. The same set of model parameters is used for multiple energy storage systems, so that the differences between PCS from different manufacturers and different control cycles are expressed through input features rather than through multi-model splitting. This forms the same decision link with the consistency set and frozen set of the consistency arrangement in the preparation stage.

[0078] The input to the effectiveness evaluation model is the unit input feature matrix, which is jointly generated from the preparation confirmation information set, version execution record, adaptive participation factor table, and changes in grid connection point power deviation. The dimension is the number of energy storage units multiplied by ten. The ten input features remain consistent for each energy storage unit, and are as follows: version consistency judgment result corresponding to the preparation confirmation information, execution direction consistency judgment result, lower limit of the executable power range corresponding to the preparation confirmation information, upper limit of the executable power range, instruction value of the energy storage unit in the submitted sub-instruction set, instruction value of the energy storage unit in the corrected sub-instruction set, set identifier of the energy storage unit in the version execution record, grid connection point power deviation, changes in grid connection point power deviation, and the current value of the energy storage unit in the adaptive participation factor table. The unit input feature matrix binds the execution boundary obtained from the two-stage submission and the net effect feedback obtained from the changes in grid connection point power deviation to the same feature chain, used to identify asynchronous distortion situations where the preparation confirmation information meets consistency requirements but no effective net effect is actually formed.

[0079] The effectiveness evaluation model consists of two fully connected layers in its intermediate layer. The first layer outputs a dimension equal to sixteen times the energy storage quantity, and the second layer outputs a dimension equal to eight times the energy storage quantity. Each fully connected layer comprises fully connected neurons. These neurons perform a weighted summation of the input features and obtain an intermediate representation through nonlinear mapping. This two-layer structure integrates the version consistency determination result, the execution direction consistency determination result, and the change in grid connection point power deviation into an interpretable effectiveness correlation representation. This enables the effectiveness evaluation metric to distinguish between versions that have been switched but whose net effect does not conform to the instruction version number. Existing practices typically treat the adaptive participation factor table as having fixed weights or communication confirmation as a reliability criterion, lacking an expression path that simultaneously incorporates the version consistency determination result and the change in grid connection point power deviation into the same structure.

[0080] The output layer of the effectiveness evaluation model is a fully connected layer. The output dimension is the number of energy storage units multiplied by one, and the output object is the effectiveness evaluation value of each energy storage unit. The effectiveness evaluation value serves as the input to S6 and is used to adjust the adaptive participation factor table. The adjustment process is driven by the effectiveness evaluation value and updates the corresponding entries in the adaptive participation factor table within the preset value boundaries.

[0081] For example, the plant-level coordinator uses the plant-level frequency modulation target. Measured power at grid connection points within the version Command version number Submit sub-instruction set Consistent set identifier vector Prepare and confirm the information set Energy storage identification vector With adaptive participation factor vector As input, generate a set of correction sub-instructions while maintaining the instruction version number. Prepare and confirm the information set. The The row includes energy storage identifiers. The corresponding lower limit of the executable power range With the upper limit of the executable power range , This refers to the amount of energy stored.

[0082] Plant-level coordinator will from The power deviation at the grid connection point is obtained by subtracting from the middle. The plant-level coordinator is based on... The sign determines the direction of the deviation. The upward direction is taken Adjust the direction downwards Take in the zero direction .

[0083] The plant-level coordinator is based on the instruction version number. Read the set of subcommit instructions ,according to Extract the command values ​​of the stored energy within the consensus set and generate a consensus set command value vector. ,in The entry with a corresponding element of 0 will Set to zero. The plant-level coordinator retrieves the prepared confirmation information set. Extract the executable power range matrix , The Travelogue The plant-level coordinator is responsible for each... Calculate the remaining upward adjustment for entries with a value of 1. With the remaining downward adjustment ,in Depend on Get and set the negative value to zero. Depend on Get and set the negative value to zero; composition ,Will composition .

[0084] Plant-level coordinator based on adaptive participation factor vector Generate assigned weight vector ,in For entries with a value of 1, take , Entries with a value of 0 are taken Plant-level coordinator Normalization yields a normalized weight vector. The normalization method is based on The sum of each element is used as the denominator for each... Perform division; when the denominator is zero, Set all elements to zero.

[0085] Plant-level coordinator based on grid connection point power deviation As the total correction amount, according to The initial correction fraction magnitude is obtained, and the initial correction fraction magnitude, deviation direction, and remaining magnitude boundary are written into the same calculation link to generate the initial correction command value vector. , of which The initial correction command value for the energy storage system is calculated using the following formula:

[0086] in, Energy storage label The initial correction instruction value; Indicates the direction of deviation; Indicates the power deviation at the grid connection point; Represents the absolute value of the power deviation at the grid connection point; Represents the normalized weight vector The One element; Indicates the remaining upward adjustment amount; Indicates the remaining downward adjustment; This indicates the operator that takes the smaller value; This indicates the operator for taking the larger value.

[0087] In this application, such as Figure 9 As shown, "intra-version deviation correction" only applies to the consistent set: the plant-level coordinator allocates intra-version errors based on the power deviation at the grid connection point, combined with the adaptive participation factor table, and outputs a "correction sub-instruction set" bound to the instruction version number. This correction link only connects to the consistent set; the frozen set is represented as "disconnected / not connected" in the diagram and does not participate in this version correction. Advantages of the invention: It avoids the mixing of frozen / asynchronous units into the closed loop, preventing the old version from overlapping and canceling out the new version, reducing correction oscillations and output distortion; at the same time, it concentrates correction resources on a controllable and verifiable set of equipment, resulting in more stable convergence.

[0088] Plant-level coordinator based on Generate executable correction instruction value vector Adjustments to be redistributed With removal of identifier vector For each For entries with a value of 1, the plant-level coordinator will Initialize to And calculate the expected magnitude of the entry. The difference between the expected amplitude and the boundary amplitude in the formula; when the expected amplitude is greater than the boundary amplitude, the difference is accumulated. And Set to 1, when the expected amplitude is not greater than the boundary amplitude. Set to 0. (This is the first step in the process.) Entries with a value of 0 will Set to zero and Setting it to 1 restricts the redistribution objects to a consistent set of energy storage that does not touch the boundaries of the executable power range.

[0089] Plant-level coordinator setting threshold for redistribution times With redistribution count ,in, For a preset integer, The initial value is 0. The plant-level coordinator is... , Non-zero and exists Perform redistribution under the following conditions: Regenerate iterative weight vectors based on entries that have not been removed. Among them, satisfying =1 and Entries The remaining entries are set to zero; Normalization yields ;Will according to The amortization yields the iterative increment vector. .when At that time, for each The entries will Add to and the remaining upward adjustment Compared to, exceeding The excess portion is accumulated back And Set as And will Set to 1; when At that time, for each The entries will Add in the negative direction until and the remaining downward adjustment Compared to, exceeding The excess portion is accumulated back And Set to negative And will Set to 1; when When Set to zero. The plant-level coordinator will Add one, and in All entries in the zero or uniform set satisfy The reallocation stops at this point, resulting in a final value that satisfies the executable power range constraint. .

[0090] The plant-level coordinator will combine the energy storage identifiers, command version numbers, and correction command values ​​of each energy storage unit within the consistency set to form a correction sub-command set. For each Energy storage generation correction sub-instruction line vector of 1 And composed in order of energy storage index This ensures that the set of correction sub-instructions and the set of commit sub-instructions maintain the same instruction version number. Furthermore, the objects to be allocated are limited to a consistent set.

[0091] In some embodiments, such as Figure 10 As shown, the step of dynamically updating the adaptive participation factor table based on the effectiveness evaluation metric to obtain the updated adaptive participation factor table includes: S1001, merge the submission sub-instruction set and the correction sub-instruction set to form the final frequency modulation sub-instruction set; S1002, based on the effectiveness evaluation quantity corresponding to each energy storage system, update the current value of the adaptive participation factor within the allowable value boundary to obtain the updated adaptive participation factor table; S1003, output the updated adaptive participation factor table and the final set of frequency modulation sub-instructions.

[0092] In this application, the set of submitted sub-instructions and the set of corrected sub-instructions are merged to form the final set of frequency modulation sub-instructions; based on the effectiveness evaluation quantity corresponding to each energy storage system, the current value of the adaptive participation factor is updated within the allowable value boundary to obtain the updated adaptive participation factor table; the updated adaptive participation factor table and the final set of frequency modulation sub-instructions are output.

[0093] For example, the plant-level coordinator uses a consistent set of identifier vectors. Modification subinstruction set Version header information vector in version execution record Prepare and confirm the information set The current value vector of the adaptive participation factor table Submit sub-instruction set As input, the process includes issuing corrective measures, generating effectiveness evaluation metrics, updating the adaptive participation factor table, and merging the frequency modulation sub-instruction set. Includes instruction version number Actual power measured at grid connection points within the version The Record the received instruction version number according to the energy storage identifier. Executable direction Lower limit of the executable power range Upper limit of the executable power range Energy storage identification vector is denoted as ,in This refers to the amount of energy stored.

[0094] Plant-level coordinator will The energy storage identifier is unicast and distributed to the consensus set. The unicast target is... The energy storage identifier with a value of 1 is determined. The plant-level coordinator records the start time of the correction and distribution. Set the duration of the correction observation window. The end time of the corrected observation window is recorded as , Through the and The result is obtained by performing the addition. The plant-level coordinator operates within the time interval. Data collection of measured power sampling matrix at grid connection point , Each line contains the sampling time. With power value ;Will The arithmetic mean of the power series is used to obtain the corrected measured power at the grid connection point after the power is distributed. The plant-level coordinator targets the plant-level frequency modulation. Based on this, the power deviation at the grid connection point before the correction was issued was calculated. Deviation from grid connection point power after correction ,in By using Deduction get, By using Deduction Obtained; Plant-level coordinator and The difference is used to obtain the change in power deviation at the grid connection point. and will The organization is for power deviation data groups .

[0095] The plant-level coordinator generates a unit input feature matrix based on one sample per energy storage unit. , Depend on Composed of feature row vectors, the first The feature row vector is denoted as . The numerical value for version consistency determination is obtained by using... and The result is obtained through comparison; a value of 1 is assigned if the two matches, and a value of 0 is assigned if they do not match. To determine the numerical value of the execution direction consistency judgment result, the following method is used: In and The command direction corresponding to energy storage The result is obtained through comparison; a value of 1 is assigned if the two matches, and a value of 0 is assigned if they do not match. Pick , Pick ; Retrieve Subcommit Set China Energy Storage Label Corresponding instruction value ,when Not included When Set to zero; Fetch Modifier Subinstruction Set China Energy Storage Label Corresponding correction instruction value ,when Not included When Set to zero; Retrieve the set identifier value from the version execution record. The consistent set is taken Freeze set ; Take the power deviation at the grid connection point , Changes in power deviation at grid connection point , Get the current value of the adaptive participation factor The plant-level coordinator will With instruction version number Bind to the feature cache to ensure sample consistency under the same instruction version number.

[0096] Plant-level coordinator will The input effectiveness evaluation model consists of a first fully connected layer, a second fully connected layer, and an output layer. The model parameters are calculated row-by-row using the same set of shared parameters for all energy storage samples. The parameters of the first fully connected layer include a weight matrix. With bias vector The input is The output is the first hidden vector. The parameters of the second fully connected layer include the weight matrix. With bias vector The input is The output is the second hidden vector. The output layer parameters include the weight matrix. With bias vector The input is The output is the effectiveness evaluation score. Both the first and second hidden vectors employ a nonlinear mapping that truncates negative values ​​to zero to form the activation result. This is used to fuse the version consistency determination result, the execution direction consistency determination result, and the change in grid connection point power deviation into a single output. The plant-level coordinator will... Organization as an effectiveness evaluation vector .

[0097] Plant-level coordinator based on Update the adaptive participation factor table and set the allowed value boundaries. and Set update step size With effective benchmark threshold For each energy storage identifier Plant-level coordinator calculation relatively The difference, when the difference is positive, is... When an adjustment is performed, the difference is negative. Perform a reduction adjustment, the adjustment magnitude being determined by comparing the difference magnitude with... Multiply to obtain; for the adjusted Perform boundary clipping to update the fall into The plant-level coordinator will be updated. The organization is the current value vector of the updated adaptive participation factor table. .

[0098] The plant-level coordinator will submit a set of sub-instructions. With the set of correction sub-instructions Merged into the final set of frequency modulation sub-instructions issued. During merging, an index is created based on the energy storage identifier, and each energy storage identifier within the consistent set is indexed. Retrieve Submission Command Value With correction instruction value Performing algebraic addition yields the merge instruction value. And retain the instruction version number. With command direction Forming frequency modulation sub-command row vector The plant-level coordinator will be all Composed of energy storage index order ,in Energy storage identifiers with a value of 0 are not written. .

[0099] like Figure 11As shown, this application centers on a "version execution record," solidifying and binding key elements related to the same frequency regulation version: instruction version number, submitted sub-instruction set, consensus set, freeze set, and the measured power of grid-connected points within that version. Each element is imported into the record through an index relationship, enabling the power response of any grid-connected point to be traced back to "which version, which energy storage units actually participated (consensus set), which were removed (freeze set), and which sub-instructions were issued at that time." This invention achieves version-level traceability and auditability, facilitating the explanation of deviation sources, locating abnormal equipment / communication problems, and avoiding the engineering pain point of "unattributable grid-connected point performance."

[0100] like Figure 12 As shown, this application provides a control device for multi-energy storage asynchronous cooperative frequency modulation output, the device comprising: The first generation module 1201 is used to generate plant-level frequency regulation targets based on the upper-level frequency regulation instructions and grid connection point power values ​​corresponding to the multiple energy storage systems connected in parallel in the thermal power plant. The power allocation module 1202 is used to determine the instruction version number based on the plant-level frequency regulation target and the preset adaptive participation factor table corresponding to each energy storage system, and to allocate power to the plant-level frequency regulation target within the executable power range of each energy storage system to obtain a set of candidate sub-instructions that are mapped one-to-one with the instruction version number; wherein, the adaptive participation factor table is used to characterize the execution reliability of each energy storage system version. The collection distribution module 1203 is used to distribute the instruction version number and the corresponding candidate sub-instruction set to each energy storage system to form a preparation request. The second generation module 1204 is used to obtain the preparation confirmation information set returned by each energy storage system in response to the preparation request, and to perform a consistency judgment on the preparation request of each energy storage system based on the preparation confirmation information set, and to generate a submit command or cancel command for the corresponding energy storage system according to the judgment result. The instruction issuing module 1205 is used to issue a submission instruction to the energy storage system corresponding to the generation of the submission command, and to issue a cancellation instruction to the energy storage system that generates the cancellation command. The power calculation module 1206 is used to calculate the grid connection point power deviation based on the plant-level frequency regulation target and the real-time acquired grid connection point power value; and to allocate the grid connection point power deviation to the energy storage system that has issued the submission instruction using the adaptive participation factor table as the allocation constraint, thereby generating a set of correction sub-instructions corresponding to the instruction version number. Instruction correction module 1207 is used to issue the correction sub-instruction set to the energy storage system within the consensus set; The third generation module 1208 is used to calculate the generation instruction execution effectiveness evaluation quantity based on the set of correction sub-instructions, the version execution record of the current instruction, the set of preparation confirmation information, and the change in the power deviation of the grid connection point. The dynamic update module 1209 is used to dynamically update the adaptive participation factor table according to the effectiveness evaluation quantity, obtain the updated adaptive participation factor table, and apply the updated adaptive participation factor table to the instruction generation and power allocation of the next plant-level frequency regulation target.

[0101] According to embodiments of this application, this application also provides an electronic device and a readable storage medium.

[0102] The electronic device includes at least one processor and a memory communicatively connected to the at least one processor. The memory stores instructions executable by the at least one processor, which, when executed, enable the at least one processor to perform the multi-energy storage asynchronous coordinated frequency modulation output control method described in this application. The computer instructions are used to cause the computer to perform the multi-energy storage asynchronous coordinated frequency modulation output control method described in this application.

[0103] This application also provides a computer program product, including a computer program / instruction, which, when executed by a processor, implements the multi-energy storage asynchronous cooperative frequency modulation output control method of this application.

[0104] Figure 13 A schematic block diagram of an example electronic device 800 that can be used to implement embodiments of this application is shown. The electronic device is intended to represent various forms of digital computers, such as laptop computers, desktop computers, workstations, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers. The electronic device may also represent various forms of mobile devices, such as personal digital processors, cellular phones, smartphones, wearable devices, and other similar computing devices. The components shown herein, their connections and relationships, and their functions are merely illustrative and are not intended to limit the implementation of the application described and / or claimed herein.

[0105] like Figure 13As shown, the electronic device 800 includes a computing unit 801, which can perform various appropriate actions and processes based on a computer program stored in a read-only memory (ROM) 802 or a computer program loaded from a storage unit 808 into a random access memory (RAM) 803. The RAM 803 may also store various programs and data required for the operation of the electronic device 800. The computing unit 801, ROM 802, and RAM 803 are interconnected via a bus 804. An input / output (I / O) interface 805 is also connected to the bus 804.

[0106] Multiple components in electronic device 800 are connected to I / O interface 805, including: input unit 806, such as keyboard, mouse, etc.; output unit 807, such as various types of displays, speakers, etc.; storage unit 808, such as disk, optical disk, etc.; and communication unit 809, such as network card, modem, wireless transceiver, etc. Communication unit 809 allows electronic device 800 to exchange information / data with other devices through computer networks such as the Internet and / or various telecommunications networks.

[0107] The computing unit 801 can be various general-purpose and / or special-purpose processing components with processing and computing capabilities. Some examples of the computing unit 801 include, but are not limited to, a central processing unit (CPU), a graphics processing unit (GPU), various special-purpose artificial intelligence (AI) computing chips, various computing units running machine learning model algorithms, a digital signal processor (DSP), and any suitable processor, controller, microcontroller, etc. The computing unit 801 performs the various methods and processes described above, such as the control method for multi-energy storage asynchronous cooperative frequency modulation output. For example, in some embodiments, the control method for multi-energy storage asynchronous cooperative frequency modulation output can be implemented as a computer software program, which is tangibly contained in a machine-readable medium, such as storage unit 808. In some embodiments, part or all of the computer program can be loaded and / or installed on the electronic device 800 via ROM 802 and / or communication unit 809. When the computer program is loaded into RAM 803 and executed by the computing unit 801, one or more steps of the control method for multi-energy storage asynchronous cooperative frequency modulation output described above can be performed. Alternatively, in other embodiments, the computing unit 801 may be configured by any other suitable means (e.g., by means of firmware) to perform a control method for multi-energy storage asynchronous coordinated frequency modulation output.

[0108] Various embodiments of the systems and techniques described above herein can be implemented in digital electronic circuit systems, integrated circuit systems, field-programmable gate arrays (FPGAs), application-specific integrated circuits (ASICs), application-specific standard products (ASSPs), systems-on-a-chip (SoCs), payload-programmable logic devices (CPLDs), computer hardware, firmware, software, and / or combinations thereof. These various embodiments may include implementations in one or more computer programs that can be executed and / or interpreted on a programmable system including at least one programmable processor, which may be a dedicated or general-purpose programmable processor, capable of receiving data and instructions from a storage system, at least one input device, and at least one output device, and transmitting data and instructions to the storage system, the at least one input device, and the at least one output device.

[0109] The program code used to implement the methods of this application may be written in any combination of one or more programming languages. This program code may be provided to a processor or controller of a general-purpose computer, special-purpose computer, or other programmable data processing device, such that when executed by the processor or controller, the functions / operations specified in the flowcharts and / or block diagrams are implemented. The program code may be executed entirely on a machine, partially on a machine, as a standalone software package partially on a machine and partially on a remote machine, or entirely on a remote machine or server.

[0110] In the context of this application, a machine-readable medium can be a tangible medium that may contain or store a program for use by or in conjunction with an instruction execution system, apparatus, or device. A machine-readable medium can be a machine-readable signal medium or a machine-readable storage medium. Machine-readable media can be, but is not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, apparatus, or devices, or any suitable combination of the foregoing. More specific examples of machine-readable storage media include electrical connections based on one or more wires, portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fibers, portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination of the foregoing.

[0111] To provide interaction with a user, the systems and techniques described herein can be implemented on a computer having: a display device for displaying information to the user (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor); and a keyboard and pointing device (e.g., a mouse or trackball) through which the user provides input to the computer. Other types of devices can also be used to provide interaction with the user; for example, feedback provided to the user can be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user can be received in any form (including sound input, voice input, or tactile input).

[0112] The systems and technologies described herein can be implemented in computing systems that include backend components (e.g., as a data server), or computing systems that include middleware components (e.g., an application server), or computing systems that include frontend components (e.g., a user computer with a graphical user interface or web browser through which a user can interact with implementations of the systems and technologies described herein), or any combination of such backend, middleware, or frontend components. The components of the system can be interconnected via digital data communication of any form or medium (e.g., a communication network). Examples of communication networks include local area networks (LANs), wide area networks (WANs), and the Internet.

[0113] Computer systems can include clients and servers. Clients and servers are generally located far apart and typically interact via communication networks. Client-server relationships are created by computer programs running on the respective computers and having a client-server relationship with each other. Servers can be cloud servers, servers in distributed systems, or servers incorporating blockchain technology.

[0114] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

Claims

1. A control method for multi-energy storage asynchronous coordinated frequency modulation output, characterized in that, The method includes: Based on the obtained upper-level frequency regulation commands and grid connection point power values ​​corresponding to multiple energy storage systems connected in parallel at thermal power plants, plant-level frequency regulation targets are generated. Based on the plant-level frequency regulation target and the preset adaptive participation factor table corresponding to each energy storage system, the instruction version number is determined. Within the executable power range of each energy storage system, the power of the plant-level frequency regulation target is allocated to obtain a set of candidate sub-instructions that are mapped one-to-one with the instruction version number. The adaptive participation factor table is used to characterize the execution reliability of each energy storage system version. The adaptive participation factor table is a pre-initialized weight table that is iteratively updated with the execution effect of the instruction. The instruction version number and the corresponding set of candidate sub-instructions are sent to each energy storage system to form a preparation request. Obtain the preparation confirmation information set returned by each energy storage system in response to the preparation request, and perform a consistency judgment on the preparation request of each energy storage system based on the preparation confirmation information set. Generate a submit command or cancel command for the corresponding energy storage system according to the judgment result. Send a submission command to the energy storage system that generates the submission command, and send a cancellation command to the energy storage system that generates the cancellation command; Based on the plant-level frequency regulation target and the real-time acquired grid connection point power value, the grid connection point power deviation is calculated; using the adaptive participation factor table as the allocation constraint, the grid connection point power deviation is allocated to the energy storage system that has issued the submission instruction, and a set of correction sub-instructions corresponding to the instruction version number is generated. The set of correction sub-instructions is issued to the energy storage systems within the consensus set; Based on the set of corrected sub-instructions, the version execution record of this instruction, the set of preparation confirmation information, and the change in the power deviation at the grid connection point, calculate and generate an instruction execution effectiveness evaluation quantity. The adaptive participation factor table is dynamically updated based on the effectiveness evaluation metric to obtain the updated adaptive participation factor table, and the updated adaptive participation factor table is applied to the instruction generation and power allocation of the next plant-level frequency regulation target.

2. The method according to claim 1, characterized in that, The generation of plant-level frequency regulation targets based on the acquired upper-level frequency regulation commands and grid connection point power values ​​corresponding to multiple parallel energy storage systems in a thermal power plant includes: Receive the frequency modulation command from the superior, extract the command arrival time and the command power target value, and use the command arrival time as the alignment reference for this frequency modulation cycle; Within this frequency modulation cycle, the grid connection point power value is collected, the grid connection point power value is time-aligned according to the alignment reference, and the grid connection point power value is averaged using a preset sampling window to obtain the grid connection point power reference value. A plant-level frequency regulation target is generated based on the commanded power target value and the grid connection point power reference value; wherein, the plant-level frequency regulation target includes: execution direction and amplitude limit; the amplitude limit is constrained by a preset plant-level adjustable power boundary.

3. The method according to claim 2, characterized in that, Based on the plant-level frequency regulation target and a preset adaptive participation factor table corresponding to each energy storage system, the instruction version number is determined. Within the executable power range of each energy storage system, power allocation is performed on the plant-level frequency regulation target to obtain a set of candidate sub-instructions that are mapped one-to-one with the instruction version number, including: The adaptive participation factor table pre-stored in the plant-level coordinator is read to obtain the corresponding entries for each energy storage system; the entries include the energy storage identifier, the current value of the adaptive participation factor, the allowed value boundary, and the reliability parameter used to characterize the reliability of version execution. The instruction boundary of the current frequency modulation cycle is determined based on the plant-level frequency modulation target, and the instruction boundary is combined with the execution direction of the plant-level frequency modulation target and the amplitude limit value to generate an instruction version number; Based on the current value of the adaptive participation factor and the plant-level frequency regulation target, initial power allocation is performed to obtain the initial decomposition command value corresponding to each energy storage system; The initial decomposition command value is limited based on the executable power range corresponding to each energy storage system. The remaining power generated by the limiting is redistributed to the energy storage systems that have not touched the boundaries of the executable power range according to the adaptive participation factor table until the decomposition commands of all energy storage systems meet the executable power range constraints. The energy storage identifier, instruction version number, instruction direction, and decomposed instruction value of each energy storage system are encapsulated into a set of candidate sub-instructions.

4. The method according to claim 1, characterized in that, The preparation request includes: Write the energy storage identifier, instruction version number, corresponding candidate sub-instruction direction, and corresponding candidate sub-instruction value into the preparation request, and configure a preparation timeout threshold for the preparation request. The preparation request is sent to each energy storage system, and a timer based on the preparation timeout threshold is started on the plant-level coordinator side.

5. The method according to claim 4, characterized in that, The process of obtaining the preparation confirmation information set returned by each energy storage system in response to the preparation request, performing a consistency judgment on the preparation request of each energy storage system based on the preparation confirmation information set, and generating a submit command or cancel command for the corresponding energy storage system according to the judgment result includes: Within the preparation timeout threshold, receive preparation confirmation information returned by each energy storage system and summarize it into a preparation confirmation information set according to the energy storage identifier; Extract the receiving instruction version number and preparation confirmation status of each energy storage system from the preparation confirmation information set, compare the receiving instruction version number with the instruction version number to obtain the version consistency determination result; Based on the comparison between the executable direction in the preparation confirmation information and the candidate sub-instruction direction of the corresponding energy storage system in the candidate sub-instruction set, the execution direction consistency determination result is obtained, and the lower limit of the executable power range and the upper limit of the executable power range are extracted from the preparation confirmation information as executable capability input; The admission weight is calculated based on the current value of the adaptive participation factor that matches the energy storage identifier in the adaptive participation factor table and the reliability parameter. The admission weight is then compared with the admission threshold to obtain the admission determination result. When the preparation confirmation status is confirmed, the version consistency determination result is consistent, the execution direction consistency determination result is consistent, and the admission determination result is passed, the corresponding energy storage identifier is included in the consistency set. If any determination result does not meet the inclusion criteria or the corresponding energy storage system fails to return preparation confirmation information within the preparation timeout threshold, the corresponding energy storage identifier will be included in the frozen set. Issue a commit command to the consistent set and an undo command to the frozen set.

6. The method according to claim 1, characterized in that, The process involves calculating the grid connection point power deviation based on the plant-level frequency regulation target and the real-time acquired grid connection point power value; using the adaptive participation factor table as an allocation constraint, allocating the grid connection point power deviation to the energy storage systems that have issued submission instructions, and generating a set of correction sub-instructions corresponding to the instruction version number, including: Based on the plant-level frequency regulation target and the real-time acquired grid connection point power value, the grid connection point power deviation is calculated, and the deviation direction of the grid connection point power deviation is determined. Read the set of submitted sub-instructions according to the instruction version number, extract the instruction value and instruction direction of each energy storage system in the consistency set, and combine the lower limit and upper limit of the executable power range of each energy storage system recorded in the preparation confirmation information set to calculate the remaining upward adjustment and remaining downward adjustment of each energy storage system relative to the current instruction value. Read the current value of the adaptive participation factor of each energy storage system in the consensus set from the adaptive participation factor table, generate the allocation weight of each energy storage system based on the current value of the adaptive participation factor, and normalize the allocation weight. Using the power deviation at the grid connection point as the total correction amount, the total correction amount is initially allocated according to the normalized allocation weight to obtain the initial correction share of each energy storage system. The initial correction fraction is combined with the deviation direction to obtain the initial correction command value for each energy storage system; The initial correction command value is compared with the remaining upward or downward adjustment range of the corresponding energy storage system at the boundary. If the initial correction instruction value exceeds the executable boundary of the corresponding energy storage system, the excess portion will be defined as the correction amount to be reallocated, and the corresponding energy storage system that touches the executable boundary will be removed from the allocation objects in subsequent dynamic adjustments. If the preset redistribution number threshold is not reached, the redistribution correction amount is recalculated and allocated to the energy storage systems that have not been removed according to the adaptive participation factor table, until the redistribution correction amount is zero, or all energy storage systems in the consensus set reach the executable power range boundary. The energy storage identifier, instruction version number, and final correction instruction value of each energy storage system in the consensus set are encapsulated to generate a correction sub-instruction set, and the correction sub-instruction set is made to have the same instruction version number as the submitted sub-instruction set.

7. The method according to claim 1, characterized in that, The step of calculating and generating an instruction execution effectiveness evaluation metric based on the set of corrected sub-instructions, the version execution record of this instruction, the set of preparation confirmation information, and the change in grid connection point power deviation includes: A set of correction sub-instructions is sent to the consensus set, the measured power of the grid connection point after the correction sub-instruction set is sent is obtained, and the change in the power deviation of the grid connection point is calculated based on the power deviation of the grid connection point before and after the correction sub-instruction set is sent. Based on the prepared confirmation information set, version execution record, adaptive participation factor table, grid connection point power deviation and the change in grid connection point power deviation, a unit input feature matrix is ​​generated for each energy storage system. The unit input feature matrix includes version consistency determination result, execution direction consistency determination result, lower limit of executable power range, upper limit of executable power range, instruction value in the submitted sub-instruction set, instruction value in the corrected sub-instruction set, set identifier in the version execution record, grid connection point power deviation, change in grid connection point power deviation, and current value of adaptive participation factor; The unit input feature matrix is ​​input into a pre-trained effectiveness evaluation model, which includes two fully connected layers and one output layer. The effectiveness evaluation value corresponding to each energy storage system is obtained by outputting the effectiveness evaluation model.

8. The method according to claim 7, characterized in that, The step of dynamically updating the adaptive participation factor table based on the effectiveness evaluation metric to obtain the updated adaptive participation factor table includes: The set of submission sub-instructions and the set of correction sub-instructions are merged to form the final set of frequency modulation sub-instructions issued. Based on the effectiveness evaluation quantity corresponding to each energy storage system, the current value of the adaptive participation factor is updated within the allowable value boundary to obtain the updated adaptive participation factor table. Output the updated adaptive participation factor table and the final set of frequency modulation sub-instructions.

9. A control device for multi-energy storage asynchronous coordinated frequency modulation output, characterized in that, The device includes: The first generation module is used to generate plant-level frequency regulation targets based on the upper-level frequency regulation instructions and grid connection point power values ​​corresponding to the multiple energy storage systems connected in parallel at the thermal power plant. The power allocation module is used to determine the instruction version number based on the plant-level frequency regulation target and the preset adaptive participation factor table corresponding to each energy storage system, and to allocate power to the plant-level frequency regulation target within the executable power range of each energy storage system to obtain a set of candidate sub-instructions that are mapped one-to-one with the instruction version number; wherein, the adaptive participation factor table is used to characterize the execution reliability of each energy storage system version. The collection distribution module is used to distribute the instruction version number and the corresponding candidate sub-instruction set to each energy storage system to form a preparation request; The second generation module is used to obtain the set of preparation confirmation information returned by each energy storage system in response to the preparation request, and to perform a consistency judgment on the preparation request of each energy storage system based on the set of preparation confirmation information, and to generate a submit command or a cancel command for the corresponding energy storage system according to the judgment result. The instruction issuing module is used to issue submission instructions to the energy storage system that generates the submission command, and to issue cancellation instructions to the energy storage system that generates the cancellation command. The power calculation module is used to calculate the grid connection point power deviation based on the plant-level frequency regulation target and the real-time acquired grid connection point power value; and to allocate the grid connection point power deviation to the energy storage system that has issued the submission instruction, using the adaptive participation factor table as the allocation constraint, and to generate a set of correction sub-instructions corresponding to the instruction version number. The instruction correction module is used to issue the correction sub-instruction set to the energy storage systems within the consensus set; The third generation module is used to calculate the evaluation value of the execution effectiveness of the generated instruction based on the set of corrected sub-instructions, the version execution record of the current instruction, the set of preparation confirmation information, and the change in the power deviation of the grid connection point. The dynamic update module is used to dynamically update the adaptive participation factor table according to the effectiveness evaluation quantity, obtain the updated adaptive participation factor table, and apply the updated adaptive participation factor table to the instruction generation and power allocation of the next plant-level frequency regulation target.

10. An electronic device, characterized in that, At least one processor; and a memory communicatively connected to the at least one processor; wherein, The memory stores instructions that can be executed by the at least one processor to enable the at least one processor to perform the method according to any one of claims 1 to 8.

Images