A micro-grid cluster coordination power distribution method based on closed-loop feedback

Abstract

The application discloses a micro-grid cluster coordination power distribution method based on closed-loop feedback, relates to the technical field of micro-grid power distribution, and comprises the following steps: according to online states and controllable states in a device state matrix, excluding offline devices from the current distribution, dividing the online and controllable devices into a controllable device set, dividing the online but uncontrollable devices into an uncontrollable device set, taking current actual operating powers of the devices in the uncontrollable device set as fixed powers, deducting the fixed powers from the corrected total target active power, and obtaining net target active power to be distributed; and the application avoids the power distribution from falling into uncontrollable or low-trust devices by combining the fixed power deduction, capacity contraction and iterative distribution, and improves the safety and executability of the device-level power control instruction.

Description

Technical Field

[0001] This invention relates to the field of microgrid power allocation technology, and in particular to a microgrid cluster coordinated power allocation method based on closed-loop feedback. Background Technology

[0002] With the large-scale integration of distributed power sources, energy storage devices, flexible loads, and electric vehicle charging and swapping equipment into the distribution network, microgrids are gradually evolving from single autonomous operation to collaborative operation within microgrid clusters. In microgrid clusters, active power is typically allocated to devices through energy management systems, edge controllers, or distributed control nodes. Existing distributed control methods often decompose the theoretical total target active power to individual distributed power sources, energy storage units, or controllable loads based on equipment rated capacity, adjustability margin, state of charge, operating costs, or preset weights to meet grid connection power constraints, load balancing, and renewable energy consumption requirements.

[0003] Microgrid clusters experience various challenges in actual operation, including changes in device online status, fluctuations in controllable status, differences in communication quality, power boundary triggering, and command response deviations. Existing coordinated power allocation methods typically focus on the static allocation of theoretical target power, neglecting the closed-loop feedback that addresses the discrepancy between the actual total active power response and the theoretical total target active power. Therefore, a closed-loop feedback-based coordinated power allocation method for microgrid clusters is needed. This method should combine device state matrices, the fixed power of uncontrollable devices, and the remaining capacity of controllable devices to achieve active power allocation more suitable for distributed control scenarios. Summary of the Invention

[0004] In view of the aforementioned existing problems, the present invention is proposed.

[0005] Therefore, this invention provides a microgrid cluster coordinated power allocation method based on closed-loop feedback to solve the problems of insufficient matching between theoretical targets and actual equipment response capabilities in the power allocation of microgrid clusters under existing distributed control scenarios, and the difficulty in effectively participating in the net target power reconstruction of fixed power of online uncontrollable equipment.

[0006] To solve the above-mentioned technical problems, the present invention provides the following technical solution:

[0007] This invention provides a microgrid cluster coordinated power allocation method based on closed-loop feedback, comprising: collecting state data of the microgrid equipment cluster; standardizing and encapsulating the state data, validating its validity, and constructing a state profile to generate an equipment state matrix, the equipment state matrix including equipment identifier, online status, controllable status, actual active power, power boundary parameters, and active power allocated in the previous cycle; calculating the deviation between the actual total response active power and the theoretical total target active power based on the actual total response active power in the equipment state matrix, and performing closed-loop correction on the theoretical total target active power to obtain the total target active power corresponding to the current power allocation cycle; and excluding offline equipment from the current allocation cycle based on the online and controllable status in the equipment state matrix. Online and controllable devices are divided into a controllable device set, and online but uncontrollable devices are divided into an uncontrollable device set. The current actual operating power of each device in the uncontrollable device set is used as the fixed power. The fixed power is subtracted from the corrected total target active power to obtain the net target active power to be allocated. Power boundary parameters and the allocated power of the controllable device set are extracted from the device state matrix. The remaining available power capacity of each controllable device is calculated. The net target active power to be allocated is iteratively allocated with boundary constraints according to the proportion of the remaining available power capacity of each controllable device to the total remaining available power capacity to obtain the target active power. The target active power of each controllable device is merged with the fixed power of each uncontrollable device to generate device-level power control instructions.

[0008] As a preferred embodiment of the microgrid cluster coordinated power allocation method based on closed-loop feedback described in this invention, the steps of collecting the status data of the microgrid equipment cluster, standardizing and encapsulating the status data, validating its validity, and constructing a status profile to generate an equipment status matrix, which includes equipment identification, online status, controllable status, actual active power, power boundary parameters, and the active power allocated in the previous cycle, are as follows:

[0009] Status data is collected from the equipment sample library and then time-aligned, unit-unified, and field-encapsulated to obtain standardized status records.

[0010] The standardized status records are validated and status profiles are constructed. The status profile values ​​are generated based on online status, communication quality, command response residuals, and power boundary validity. The online status, controllable status, power boundary, and actual power are integrated according to the device identifier to generate a device status matrix.

[0011] As a preferred embodiment of the microgrid cluster coordinated power allocation method based on closed-loop feedback described in this invention, the specific steps for calculating the deviation between the actual total response active power and the theoretical total target active power based on the actual total response active power in the device state matrix are as follows:

[0012] Power, status, communication, and boundary data are read from the device status matrix and device sample library, and the device response sequence is obtained by filtering.

[0013] Construct an electrical-communication interaction diagram for the equipment, perform consistency correction and residual filtering on the equipment response sequence, and generate deviation values, deviation levels, and abnormal equipment identifiers.

[0014] As a preferred embodiment of the microgrid cluster coordinated power allocation method based on closed-loop feedback described in this invention, the specific steps for obtaining the total target active power corresponding to the current power allocation cycle are as follows:

[0015] Based on the deviation value, deviation level, and abnormal equipment identifier, combined with the total target active power corrected in the previous cycle, the current theoretical total target power, and the power carrying capacity boundary, a closed-loop correction trigger quantity is generated.

[0016] The total target power correction constraint domain is formed based on the closed-loop correction trigger value. Within the total target power correction constraint domain, the theoretical total target power is recursively corrected based on the historical deviation trend, the continuity of the correction direction, and the equipment reliability status to obtain the total target active power corresponding to the current power allocation cycle.

[0017] As a preferred embodiment of the microgrid cluster coordinated power allocation method based on closed-loop feedback described in this invention, the device electrical-communication interaction graph is constructed by establishing electrical connection edges based on each device as a node, according to the bus affiliation and feeder connection relationship in the device sample library, determining the communication reliability based on the heartbeat, communication delay and packet loss rate in the network status, and then adjusting the node weights.

[0018] As a preferred embodiment of the microgrid cluster coordinated power allocation method based on closed-loop feedback described in this invention, the steps of excluding offline devices from the current allocation based on the online and controllable states in the device state matrix, dividing online and controllable devices into a controllable device set, dividing online but uncontrollable devices into an uncontrollable device set, and using the current actual operating power of each device in the uncontrollable device set as the fixed power are as follows:

[0019] State feature data is read from the device state matrix, and adjacency relationships and edge reliability coefficients are extracted by combining the device electrical-communication interaction graph to obtain the basic data for device clustering.

[0020] Based on the basic data of device clustering, basic state trust features, communication trust features, residual trust features, neighborhood consistency features and power boundary security features are generated to form a core feature group. The core feature group is then weighted and normalized to obtain the allocatable latch value for this round.

[0021] If a device is offline, it will be excluded from this round of allocation.

[0022] If a device is online and controllable, and the allocatable latch value in this round reaches the latch threshold, it is classified as a controllable device set.

[0023] If a device is online but uncontrollable, or if a device is online and controllable but the available latch value for this round has not reached the latch threshold, it is classified as an uncontrollable device set.

[0024] The current actual active power of each device in the set of uncontrollable devices is locked as the fixed power, and the device identifier, fixed power value and fixed reason are stored accordingly.

[0025] As a preferred embodiment of the microgrid cluster coordinated power allocation method based on closed-loop feedback described in this invention, the specific steps for subtracting the fixed power from the corrected total target active power to obtain the net target active power to be allocated are as follows:

[0026] Read the corrected total target power and the uncontrollable equipment curing power, and perform risk equivalence processing on the curing power to obtain the equivalent curing power;

[0027] If the set of controllable devices is empty, the power allocation of controllable devices in this round is stopped, and the fixed power of uncontrollable devices is used as the output power command for this round. If the set of controllable devices is not empty, the equivalent fixed power is deducted and the boundary is limited to obtain the net target power to be allocated.

[0028] As a preferred embodiment of the microgrid cluster coordinated power allocation method based on closed-loop feedback described in this invention, the specific steps for extracting power boundary parameters and the allocated power of the controllable device set from the device state matrix, and calculating the remaining available power capacity of each controllable device are as follows:

[0029] Extract the power boundary, allocated power, actual power, and status profile values ​​of controllable devices from the device status matrix to generate a capacity ledger;

[0030] By combining boundary margin, ramp achievable amount, execution deviation and neighborhood consistency, the remaining available power capacity of each controllable device is calculated;

[0031] The remaining available power capacity is reduced according to the shrinkage coefficient corresponding to the status profile value, execution deviation, and neighborhood risk value. For devices with no remaining available power capacity, they will not participate in the subsequent capacity allocation, and a remaining capacity table will be generated.

[0032] As a preferred embodiment of the microgrid cluster coordinated power allocation method based on closed-loop feedback described in this invention, the specific steps for obtaining the target active power are as follows:

[0033] Read the net target active power to be allocated, the remaining capacity table, the power boundary of controllable equipment, the allocated power, the status profile value, and the equipment electrical-communication interaction diagram, and establish an iterative allocation ledger;

[0034] The initial allocation share is generated based on the percentage of remaining capacity according to the iterative allocation ledger, and the share is suppressed for low-confidence devices to obtain the candidate power allocation amount.

[0035] In each round, the candidate power allocation is written back to the iterative allocation ledger, the edge devices are locked, the unallocated residuals are fed back, and the target active power is generated after the convergence condition is met. Devices that have reached the upper and lower power limits are removed from the subsequent allocation set, and the unallocated residuals are redistributed among the unlocked devices with remaining capacity according to the proportion of remaining capacity, until the residuals are less than the preset convergence threshold.

[0036] As a preferred embodiment of the microgrid cluster coordinated power allocation method based on closed-loop feedback described in this invention, the specific steps for generating device-level power control commands are as follows:

[0037] Read the target power of controllable devices and the fixed power of uncontrollable devices from the iterative allocation ledger and the list of fixed power of uncontrollable devices, and establish a device-level instruction ledger;

[0038] The merged power is subjected to boundary verification, step size shaping and instruction type marking, and power dispatch instructions are encapsulated based on the instruction ledger to generate device-level power control instructions.

[0039] The beneficial effects of this invention are as follows: By using the electrical-communication interaction diagram of the equipment and the deviation closed-loop correction, the microgrid cluster under distributed control can dynamically correct the total target power by combining the actual response, communication reliability status and electrical adjacency relationship, thereby improving the scheduling accuracy; by dividing the controllable equipment set and the uncontrollable equipment set by the allocable latch value in this round, and by combining fixed power deduction, capacity shrinkage and iterative allocation, power allocation is avoided from falling into uncontrollable or low-reliability equipment, thereby improving the security and executability of equipment-level power control commands. Attached Figure Description

[0040] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the following description of the embodiments will be briefly introduced. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0041] Figure 1 This is a flowchart of a microgrid cluster coordinated power allocation method based on closed-loop feedback.

[0042] Figure 2A flowchart for device grouping and power locking.

[0043] Figure 3 This is a flowchart for generating a table of remaining capacity.

[0044] Figure 4 A flowchart for generating the target active power. Detailed Implementation

[0045] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, the specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings.

[0046] Many specific details are set forth in the following description in order to provide a full understanding of the invention. However, the invention may also be practiced in other ways different from those described herein, and those skilled in the art can make similar extensions without departing from the spirit of the invention. Therefore, the invention is not limited to the specific embodiments disclosed below.

[0047] Secondly, the term "one embodiment" or "embodiment" as used herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The phrase "in one embodiment" appearing in different places in this specification does not necessarily refer to the same embodiment, nor is it a single or selective embodiment that is mutually exclusive with other embodiments.

[0048] Reference Figures 1-4 This is one embodiment of the present invention, which provides a microgrid cluster coordinated power allocation method based on closed-loop feedback, comprising the following steps:

[0049] S1. Collect status data of microgrid equipment clusters, standardize and encapsulate the status data, verify its validity, and construct status profiles to generate an equipment status matrix. The equipment status matrix includes equipment identification, online status, controllable status, actual active power, power boundary parameters, and active power allocated in the previous cycle.

[0050] S1.1 Collect status data from the equipment sample library, and perform time alignment, unit unification, and field encapsulation to obtain standardized status records.

[0051] Furthermore, real-time status data of microgrid devices is read from the device acquisition terminal, edge controller, energy management system or electrical monitoring system, and basic data such as device rated parameters, power boundaries, unique device identifiers, bus affiliation, feeder connection relationships and unit mapping relationships are read from the device sample library; the acquisition time of real-time status data is uniformly converted into UTC millisecond timestamps and aligned to the standard time index table in one-second cycles;

[0052] When the time deviation between the collection timestamp of the status field and the corresponding standard time index exceeds 500 milliseconds, the status field is marked as missing; for status fields that do not exceed the time deviation, the unit is converted according to the status field unit mapping table, and the status field is encapsulated according to the device unique identifier and standard timestamp to generate a standardized status record.

[0053] It should be noted that status data includes device identification data, time data, operating power data, control status data, response deviation data, and neighborhood communication data;

[0054] The basic data in the equipment sample library includes the equipment's rated power, minimum allowable power, maximum allowable power, adjustment step size, ramp limit, bus assignment, feeder connection relationship, and field unit mapping relationship.

[0055] S1.2 Validate the standardized status records and construct status profiles. The status profile values ​​are generated based on online status, communication quality, command response residuals and power boundary validity. The online status, controllable status, power boundary and actual power are integrated according to the device identifier to generate the device status matrix.

[0056] Furthermore, for each standardized status record, missing markers, acquisition quality indicators, unit anomaly markers, and standard status values ​​are read, and the validity of each status field is determined according to preset verification conditions. When a status field meets the verification conditions, it is marked as a valid status field; when a status field does not meet the verification conditions, it is marked as an invalid status field. Based on the valid status fields, the online status, controllable status, communication quality, command response residual, actual active power, minimum allowable power, and maximum allowable power of the device are extracted, and these fields are used as the basic data for constructing the status profile.

[0057] A status profile value is generated based on the device's online status, communication quality, command response residuals, and power boundary validity. When the device is online, communication quality is good, command response residuals are within the allowable range, and actual active power is within the power boundary, the status profile value is increased. When the device is offline, communication quality is abnormal, command response residuals exceed the allowable range, power boundaries are invalid, or actual active power exceeds the boundary, the status profile value is decreased. The online status, controllable status, status profile value, power boundary parameters, actual active power, and the active power allocated in the previous cycle are integrated according to the device's unique identifier and standard timestamp to generate a device status matrix.

[0058] It should be noted that state profile construction refers to converting the device's online status, communication quality, command response residuals, and power boundary validity into state profile values ​​that characterize the device's current operational credibility and its ability to participate in power allocation. The state profile values ​​are represented by normalized numerical values, ranging from 0 to 1. The larger the value, the more suitable the device is to participate in the current round of power allocation; the smaller the value, the less suitable the device is to participate in the current round of power allocation. When the state profile value is lower than the preset state profile threshold, the corresponding device is marked as a low-confidence device.

[0059] The verification conditions include missing data being marked as not missing, data acquisition quality being marked as valid, unit anomalies being marked as not anomalies, and the standard status value being within the range specified in the status field value range table. Status fields that do not meet any of the verification conditions are marked as invalid status fields.

[0060] S2. Based on the actual total response active power in the equipment state matrix, calculate the deviation between the actual total response active power and the theoretical total target active power, and perform closed-loop correction on the theoretical total target active power to obtain the total target active power corresponding to the current power allocation cycle.

[0061] S2.1 Read power, status, communication and boundary data from the device status matrix and device sample library, and filter to obtain device response sequences.

[0062] Furthermore, the system reads the device's unique identifier, standard timestamp, online status, controllable status, and actual active power from the device status matrix, and reads or supplements the minimum allowable power, maximum allowable power, communication status, and acquisition quality identifier from the device sample library. Data that meets the screening criteria is sorted in ascending order by device unique identifier and standard timestamp, and continuous time point data is retained to generate a device response sequence. The device response sequence includes the actual power data of online controllable devices and the actual power data of online uncontrollable devices.

[0063] It should be noted that the screening criteria are: online status is 1, communication status is normal, data acquisition quality identifier is valid, and actual active power is valid. Among them, valid actual active power means that the actual active power is within the range of minimum allowable power to maximum allowable power. Controllable status is not a necessary condition for screening the device response sequence in this step, but is reserved as a field in the device response sequence for subsequent division of controllable device set and uncontrollable device set.

[0064] S2.2 Construct the electrical-communication interaction diagram of the equipment, and perform consistency correction and residual filtering on the equipment response sequence to generate deviation values, deviation levels and abnormal equipment identifiers.

[0065] Furthermore, using the unique identifier of each device as a graph node, the system reads the busbar affiliation, feeder number, grid connection point number, switch topology, and adjacent feeder connection relationships of each device from the device sample library. When two devices belong to the same busbar, adjacent sections of the same feeder, or are electrically connected through a closed switch, an electrical connection edge is established between the corresponding two device nodes, and the connection edge type, the busbar to which it belongs, the feeder to which it belongs, and the valid electrical connection marker are recorded. Simultaneously, the system reads the heartbeat status, communication latency, packet loss rate, and data update time difference of each device from the network status data. When two devices share a communication gateway, are on the same communication link, or have a collaborative control relationship, a communication connection edge is established between the corresponding two device nodes. The communication reliability is determined based on the heartbeat status, communication latency, packet loss rate, and data update time difference. The electrical connection edge and the communication connection edge are merged, and the node weight and edge reliability coefficient are adjusted according to the communication reliability to obtain the device electrical-communication interaction graph.

[0066] The device response sequence is sorted by device unique identifier and standard timestamp, and matched with the adjacency relationship in the device electrical-communication interaction diagram. During consistency correction, records with abnormal time intervals, abnormal online status, abnormal communication status, or actual active power exceeding the limit are removed. The actual active power of adjacent devices under the same standard timestamp is read, and the median of the actual active power of adjacent devices is used as the reference power. Response values ​​with single-point mutations, missing additions, or inconsistent trends with the neighborhood are corrected and marked to obtain the corrected device response sequence. The difference between the actual active power of a single device and the corresponding reference power is used as the device response residual. A residual rate is generated based on the device response residual. Instantaneous residuals that do not reach the residual level threshold or only appear at a single time point are filtered out. Device records that reach level one residual or above for three consecutive standard timestamps are retained, and abnormal device identifiers and corresponding residual levels are generated. At the same time, the corrected actual active power of each online device under the same standard timestamp is summed to obtain the actual total response active power. The actual total response active power is compared with the current theoretical total target active power to generate the total power deviation for closed-loop correction.

[0067] It should be noted that the equipment response residual level threshold is set based on the rated active power of the corresponding equipment and is used to judge the degree of response abnormality of a single equipment; the total power deviation level threshold is set based on the total rated active power of the microgrid equipment cluster or the current theoretical total target active power and is used to judge the degree of overall response deviation of the microgrid equipment cluster.

[0068] Both the equipment response residual level threshold and the total power deviation level threshold can be preset based on the acquisition accuracy, normal residual fluctuation, neighborhood power dispersion, and scheduling allowable error. Typically, the first-level threshold is determined through historical normal operating condition statistics, and the second-level and third-level thresholds are divided according to safety margin. Among them, the exemplary values ​​of the equipment response residual level threshold can be 3% to 8%, 8% to 15%, and 15% to 30% of the rated active power of the corresponding equipment, and the exemplary values ​​of the total power deviation level threshold can be 2% to 5%, 5% to 10%, and 10% to 20% of the total rated active power of the cluster or the current theoretical total target active power.

[0069] S2.3. Based on the deviation value, deviation level, and abnormal equipment identifier, combined with the total target active power corrected in the previous cycle, the current theoretical total target power, and the power carrying capacity boundary, generate the closed-loop correction trigger quantity.

[0070] Furthermore, within the current control cycle, the total power deviation, deviation level, and abnormal equipment identifier are read, and the closed-loop correction direction is determined based on the direction of the total power deviation. When the actual total response active power is lower than the current theoretical total target power, it is determined to be insufficient response, and an abnormal deviation compensation amount is generated in the direction of increasing the target power. When the actual total response active power is higher than the current theoretical total target power, it is determined to be excessive response, and an abnormal deviation compensation amount is generated in the direction of decreasing the target power. For the equipment corresponding to the abnormal equipment identifier, the magnitude of the abnormal deviation compensation amount is determined according to the deviation level weight, where the weight of the first-level deviation is 0.5, the weight of the second-level deviation is 0.8, and the weight of the third-level deviation is 1.

[0071] The abnormal deviation compensation amount is applied to the current theoretical total target power according to the closed-loop correction direction to obtain the initial correction target power. If the initial correction target power is lower than the minimum borne power, the minimum borne power is taken as the correction target power. If the initial correction target power is higher than the maximum borne power, the maximum borne power is taken as the correction target power. If the initial correction target power is within the range of the minimum borne power to the maximum borne power, the initial correction target power is retained as the correction target power. The difference between the correction target power and the total target active power corrected in the previous cycle is used to generate the closed-loop correction trigger amount.

[0072] S2.4. Based on the closed-loop correction trigger quantity, a total target power correction constraint domain is formed. Within the total target power correction constraint domain, the theoretical total target power is recursively corrected based on the historical deviation trend, the continuity of the correction direction, and the equipment reliability status to obtain the total target active power corresponding to the current power allocation cycle.

[0073] It should be noted that the lower limit of the total target power correction constraint domain is the current theoretical total target power minus the absolute value of the closed-loop correction triggering amount, and the upper limit of the constraint is the current theoretical total target power plus the absolute value of the closed-loop correction triggering amount.

[0074] The initial value, preset upper limit, preset lower limit, preset gain amplitude, and preset attenuation amplitude of the correction coefficient can be preset or calibrated according to the microgrid equipment response speed, power regulation accuracy, and historical deviation fluctuations. In one example, the initial value of the correction coefficient can be 0.5, the preset lower limit can be 0.2, the preset upper limit can be 1, and it increases by 0.1 per cycle when the correction direction is the same, and decreases by 0.2 per cycle when the correction direction is opposite.

[0075] Furthermore, using the total target active power corrected in the previous cycle as a recursive benchmark, the current closed-loop correction trigger value is extracted, and the correction direction is determined by combining the average total power deviation of the last three control cycles. The correction coefficient has a preset initial value, a preset upper limit, and a preset lower limit. When the current correction direction is consistent with the correction direction of the previous cycle, the correction coefficient is increased by a preset gain, and the increased correction coefficient does not exceed the preset upper limit. When the current correction direction is opposite to the correction direction of the previous cycle, the correction coefficient is decreased by a preset attenuation, and the decreased correction coefficient is not lower than the preset lower limit. When the current closed-loop correction trigger value is lower than the preset trigger threshold, the correction coefficient remains unchanged to avoid target power oscillations caused by small deviations. The device reliability coefficient is determined based on the proportion of reliable devices to controllable devices; the higher the proportion of reliable devices, the larger the allowable correction range.

[0076] S3. Based on the online and controllable states in the device status matrix, offline devices are excluded from this round of allocation. Online and controllable devices are divided into a controllable device set, and online but uncontrollable devices are divided into an uncontrollable device set. The current actual operating power of each device in the uncontrollable device set is used as the fixed power. The fixed power is deducted from the corrected total target active power to obtain the net target active power to be allocated.

[0077] S3.1 Read the state feature data from the equipment state matrix, and extract the adjacency relationship and edge reliability coefficient by combining the equipment electrical-communication interaction graph to obtain the basic data for equipment grouping.

[0078] Furthermore, using the current standard timestamp as an index, the device identifier is read row by row from the device status matrix, and the online status, controllable status, current actual active power, status profile value, communication quality, command response residual, minimum allowable power, and maximum allowable power are extracted synchronously according to the device identifier to form a status entry for a single device; using the device identifier as the query key, the electrical connection edge and communication connection edge directly connected to the device are retrieved in the device electrical-communication interaction diagram, and the device identifier corresponding to the other end of the connection edge is registered as the adjacent device identifier to form an adjacency relationship;

[0079] For each adjacency relationship, the electrical connection validity flag, communication link quality, data update time difference, and state consistency flag corresponding to the connection edge are read, and the edge reliability coefficient is determined based on the degree of connection validity, communication stability, data not timed out, and consistent state changes. The device identifier, single device state entry, adjacent device identifier, connection edge type, and edge reliability coefficient are bound according to the device identifier to obtain the basic data for device clustering.

[0080] It should be noted that the status characteristic data includes device identification, online status, controllable status, current actual active power, status profile value, communication quality, command response residual, minimum allowable power, and maximum allowable power. Among them, online status and controllable status are used to initially determine whether the device is qualified for allocation, communication quality, command response residual, and status profile value are used to determine the reliability of the device operation, and minimum allowable power and maximum allowable power are used to determine the power adjustment boundary of the device.

[0081] S3.2. Based on the device clustering basic data, generate basic state trust features, communication trust features, residual trust features, neighborhood consistency features and power boundary security features to form a core feature group. Then, perform weighted normalization on the core feature group to obtain the allocatable latch value for this round.

[0082] Furthermore, based on the basic data of device clustering, device identifiers are extracted one by one, and the corresponding online status, controllable status, and status profile values ​​are called to verify whether the device has the basic operating conditions to participate in this round of allocation, generating basic status credibility features; communication quality and edge credibility coefficients related to the device are extracted to verify whether the device's own communication status and adjacent connections are stable, generating communication credibility features; command response residuals are extracted and converted into residual credibility features, where the smaller the command response residual, the higher the residual credibility feature, and the larger the command response residual, the lower the residual credibility feature; based on the adjacency relationship, the status profile values ​​and current actual active power of adjacent devices are reviewed, and combined with the edge credibility coefficients to determine whether the status and power changes between the current device and adjacent devices are coordinated, generating neighborhood consistency features;

[0083] The distance between the current actual active power and the power boundary is converted into a power boundary security feature. The further the current actual active power is from the power boundary, the higher the power boundary security feature; the closer the current actual active power is to the power boundary or if it exceeds the boundary, the lower the power boundary security feature. The basic state credibility feature, communication credibility feature, residual credibility feature, neighborhood consistency feature, and power boundary security feature are combined into a core feature group according to the device identifier, and normalized according to the preset weights to obtain the current round allocable latch value with a value range of 0 to 1. The larger the latch value, the more suitable the device is to participate in the current round of power allocation.

[0084] The expression for calculating the allocatable latch value in this round is as follows:

[0085] ;

[0086] in, Indicates device identification The corresponding allocatable latch value for this round; Indicates the device identifier currently to be judged; Indication and equipment Identifiers of adjacent devices that have a mutual adjacency relationship; Indicates device The set of adjacent devices; Indicates device and adjacent devices The edge reliability coefficient of the connecting edges is jointly set by electrical connection validity, heartbeat status, communication latency, packet loss rate and data update time difference; Indicates device The neighborhood normalization coefficient is used to normalize the confidence coefficient of adjacent edges. It is obtained by summing the confidence coefficients of the edges between the current device and all valid adjacent devices. Indicates device The basic state credibility characteristics, Indicates state; Indicates device The trustworthy characteristics of communication means communication; Indicates device The reliability characteristics of the residuals Represents the residual; Indicates device The power boundary safety features, Indicates the boundary; Indicates device Relative to adjacent devices The neighborhood consistency characteristic is determined based on the edge reliability coefficients between the current device and its neighboring devices. Indicates the neighborhood.

[0087] It should be noted that the core feature group is a multi-dimensional evaluation set formed by combining features such as basic state reliability, communication reliability, residual reliability, neighborhood consistency, and power boundary security, using the device identifier as an index. It is used to comprehensively determine whether the device is suitable to participate in this round of power allocation.

[0088] All features participating in this round of allocable latch value calculation are positive features ranging from 0 to 1, with larger values ​​indicating that the device is more suitable for allocation. Among them, the state reliability feature does not repeat and includes communication quality, command response residual, and power boundary proximity. Communication quality is converted into communication reliability feature, command response residual is converted into residual reliability feature, and power boundary proximity is converted into power boundary security feature. The neighborhood consistency feature is used to characterize the consistency of the state and power changes of the current device and its neighboring devices. The neighborhood normalization coefficient is the sum of the edge reliability coefficients between the current device and all its neighboring devices. When there are no effective neighboring devices, a preset default value is used.

[0089] The allocation conditions refer to the equipment being online, controllable, having reliable communication, normal command response residuals, consistent neighborhood status, and the current actual active power not being close to the power boundary.

[0090] S3.3 If the device is offline, it will be excluded from this round of allocation.

[0091] Furthermore, based on the basic data of device clustering, the device identifier is read one by one, and the online status of the corresponding device is checked first. When the online status of a device is offline, the device is determined to not have the basic access conditions for this round of power allocation, and the controllable status, communication quality, command response residual, neighborhood consistency characteristics, and power boundary security characteristics are no longer checked. At the same time, the device identifier is written into the exclusion record of this round, and the reason for exclusion is recorded as offline status, so that the device does not enter the subsequent division process of controllable device set and uncontrollable device set, and is directly removed from the allocation objects of this round.

[0092] S3.4 If the device is online and controllable, and the allocatable latch value in this round reaches the latch threshold, then it is classified as a controllable device set.

[0093] Furthermore, based on the basic data of device clustering, the device identifier is read one by one. After excluding offline devices, the online status, controllable status, and current-round allocatable latch value of the corresponding device are checked. When the device is online and controllable, it is considered a candidate controllable device, and the current-round allocatable latch value and latch threshold of the device are read. If the current-round allocatable latch value is greater than or equal to the latch threshold, the device is determined to have met the stable allocation requirements for this round, and the device identifier is written into the controllable device set. If the current-round allocatable latch value is less than the latch threshold, the device is determined to have not yet stably met the allocation requirements for this round, is not written into the controllable device set, and enters the subsequent division of the uncontrollable device set.

[0094] It should be noted that the latch threshold can be preset based on communication stability, response success rate and historical allocation success rate, and determined by the distribution of latch values ​​under normal and controllable working conditions. An exemplary value is 60% to 90% of the full-scale latch value, preferably 70% to 85%.

[0095] S3.5 If a device is online but uncontrollable, or if a device is online and controllable but the available latch value for this round has not reached the latch threshold, it is classified as an uncontrollable device set.

[0096] Furthermore, the online status, controllable status, and allocable latch value for each device to be assigned are read from the device status determination table. Whether the device has real-time scheduling capability is used as the basis for classifying it as uncontrollable. If the device is online but the controllable status is marked as uncontrollable, it means that although the device maintains a communication connection, it does not currently have the ability to execute control commands, and the device is directly marked as an uncontrollable device. If the device is online and the controllable status is marked as controllable, the allocable latch value for the device in this round of scheduling is further checked to see if it reaches the preset latch threshold.

[0097] When the allocatable latch value in this round is less than the latch threshold, it is determined that although the device has the basic online and controllable conditions, the current allocation stability is insufficient and does not meet the control allocation requirements of this round. The device is then marked as a device with insufficient latch. The device identifier, online status, controllable status, allocatable latch value in this round, latch threshold, and reason for uncontrollability are uniformly registered for both uncontrollable and insufficient latch devices, and a set of uncontrollable devices is generated.

[0098] S3.6 Lock the current actual active power of each device in the uncontrollable device set as the fixed power, and store the device identifier, fixed power value and fixed reason accordingly.

[0099] Furthermore, based on the set of uncontrollable devices, the device identifier, current actual active power, and the reason for uncontrollability are read one by one. The current actual output state of the device is used as the fixed boundary in subsequent scheduling calculations. For each uncontrollable device, the target active power for this round is not recalculated, nor is it involved in the allocation of adjustable power margin. Instead, the current actual active power is directly extracted as a fixed power value.

[0100] If a device enters the uncontrollable device set because its controllable state has changed from uncontrollable, its current actual active power is locked as the uncontrollable state fixed power. If a device enters the uncontrollable device set because its allocable latch value in this round has not reached the latch threshold, its current actual active power is locked as insufficient latch fixed power. For locked devices, their device identifier, current actual active power, fixed power value, reason for locking, and corresponding uncontrollable source are retained, and a one-to-one correspondence between device identifiers and fixed power values ​​is established to form an uncontrollable device fixed power record.

[0101] It should be noted that the fixed power represents the power value of the devices in the set of uncontrollable devices that will not be adjusted in this round of allocation; the reasons for fixed power include the controllable state being uncontrollable, or the allocable latch value in this round not reaching the latch threshold.

[0102] S3.7 Read the corrected total target power and the uncontrollable equipment solidification power, and perform risk equivalence processing on the solidification power to obtain the equivalent solidification power.

[0103] Furthermore, the total target power after this round of correction is obtained, and the equipment identifier, fixed power value, and fixed cause are retrieved from the fixed power record of uncontrollable equipment. The actual output locked by the uncontrollable equipment is used as the unadjustable occupancy in the total target power allocation. The equivalent method is distinguished according to the fixed cause. If the fixed cause is due to the uncontrollable state of the equipment, the fixed power is regarded as the unrecoverable occupancy power and is equivalently processed according to the preset state risk coefficient. If the fixed cause is due to the allocable latch value in this round not reaching the latch threshold, the fixed power is regarded as the occupancy power due to insufficient latch stability and is equivalently processed according to the preset latch risk coefficient.

[0104] When there are fluctuation records, state switching records, or historical response deviation records in the solidified power value, a risk compensation amount is introduced to correct the solidified power. The risk compensation amount is determined based on the degree of solidified power fluctuation, the number of state switching times, and the degree of historical response deviation; the greater the solidified power fluctuation, the more frequent the state switching, or the greater the historical response deviation, the greater the risk compensation amount. The power value after risk coefficient processing and risk compensation processing is marked as the equivalent solidified power, and associated with the equipment identifier, original solidified power value, solidification reason, risk coefficient, risk compensation amount, and equivalent solidified power value, generating an equivalent solidified power record. This record is used to subsequently deduct the risk-occupied portion of uncontrollable equipment from the corrected total target power.

[0105] It should be noted that when the cause of solidification changes from a controllable state to an uncontrollable state, the state risk coefficient is one or greater than one. When the state risk coefficient is one, the equivalent solidification power equals the solidification power; when the state risk coefficient is greater than one, the equivalent solidification power is greater than the solidification power. If the cause of solidification is that the allocatable latch value in this round has not reached the latch threshold, the latch risk coefficient is determined based on the difference between the latch value and the latch threshold; the larger the difference, the larger the latch risk coefficient.

[0106] S3.8 If the set of controllable devices is empty, stop the power allocation of controllable devices in this round and use the fixed power of uncontrollable devices as the output power command for this round. If the set of controllable devices is not empty, deduct the equivalent fixed power and limit the boundary to obtain the net target power to be allocated.

[0107] Furthermore, based on the set of controllable devices, the number of controllable devices, device identifiers, and their power capacity boundaries for this round are determined, and the corrected total target power and equivalent solidified power records are obtained simultaneously. If the set of controllable devices is empty, it is determined that there are no devices to receive the remaining power in this round, the power allocation of controllable devices in this round is stopped, the net target power to be allocated is set to zero, and an alarm for no controllable devices is generated. At the same time, the solidified power of the uncontrollable devices is recorded as the solidified power output record for this round.

[0108] If the set of controllable devices is not empty, the equivalent solidified power of each uncontrollable device is summed to obtain the total equivalent solidified power. The total equivalent solidified power is then subtracted from the corrected total target power to obtain the initial net target power to be allocated. A limiting boundary is formed based on the minimum and maximum achievable power of the set of controllable devices. If the initial net target power to be allocated is lower than the lower limit, it is corrected to the lower limit. If it is higher than the upper limit, it is corrected to the upper limit. If it is within the limiting boundary, it remains unchanged. The net target power to be allocated is obtained, and the status of the set of controllable devices, the total equivalent solidified power, the limiting boundary, the processing reason, and the net target power value to be allocated are associated and saved.

[0109] S4. Extract the power boundary parameters and the allocated power of the controllable device set from the device state matrix, calculate the remaining available power capacity of each controllable device, and perform iterative allocation with boundary constraints on the net target active power to be allocated according to the proportion of the remaining available power capacity of each controllable device to the total remaining available power capacity, so as to obtain the target active power.

[0110] S4.1 Extract the power boundary, allocated power, actual power, and status profile value of controllable devices from the device status matrix to generate a capacity ledger.

[0111] Furthermore, based on the set of controllable devices, the device identifier of each controllable device is determined, and the corresponding upper power limit, lower power limit, actual power, status profile value, and ramp limit value are extracted from the device status matrix. At the same time, the power allocated to the corresponding device in the previous cycle is retrieved from the device-level power control command or historical allocation record of the previous cycle. The power boundary is used as the basis for capacity range, the power allocated in the previous cycle and the current actual power are used as the basis for capacity occupancy and execution deviation verification, and the status profile value and ramp limit value are used as the basis for subsequent adjustable capacity constraints.

[0112] If the actual power of a device is close to the power boundary, it is marked as boundary-restricted; if the power allocated in the previous cycle deviates from the current actual power, it is marked as allocation deviation; if the status profile value or ramp limit does not meet the allocation requirements of this round, it is marked as status ramp-restricted; for each device, the device identifier, power boundary, power allocated in the previous cycle, current actual power, status profile value, ramp limit, constraint source, and capacity judgment result are associated and saved to generate a capacity ledger.

[0113] It should be noted that the capacity ledger is used to record the power boundaries, power occupancy, operating status, and adjustability of controllable devices before they participate in power allocation.

[0114] S4.2. Combine boundary margin, ramp achievable amount, execution deviation and neighborhood consistency to calculate the remaining available power capacity of each controllable device.

[0115] Furthermore, based on the capacity ledger, the device identifier, power boundary, current actual power, ramp limit, previously allocated power, execution deviation, and neighborhood consistency results for each controllable device are determined. Boundary margins are determined according to the adjustment direction of the net target power to be allocated in this round. When the current round is in the power increase direction, the difference between the maximum allowable power and the current actual power is used as the boundary margin; when the current round is in the power decrease direction, the difference between the current actual power and the minimum allowable power is used as the boundary margin. The ramp achievable capacity is determined based on the device ramp limit and the current control cycle length, and the smaller value between the boundary margin and the ramp achievable capacity is used as the basic available capacity.

[0116] The basic available capacity is adjusted based on the status profile value, execution deviation, and neighborhood consistency results. The lower the status profile value, the greater the capacity reduction; the greater the execution deviation, the greater the capacity reduction; and the lower the neighborhood consistency, the greater the capacity reduction. Boundary checks are performed on the adjusted capacity results. If the adjusted capacity is less than zero, it is set to zero and marked as unavailable. If the adjusted capacity exceeds the device's allowable capacity boundary, it is limited according to the device's allowable capacity boundary. For each device, the device identifier, boundary margin, ramp-up achievable amount, execution deviation, neighborhood consistency result, reason for capacity adjustment, and remaining available power capacity value are associated and saved, generating a record of the remaining capacity of controllable devices.

[0117] It should be noted that the remaining available power capacity is based on the smaller of the boundary margin and the ramp-up achievable amount, and is obtained after being reduced according to the state profile value, the degree of execution deviation, and the neighborhood consistency results. It is used to represent the amount of power regulation that the equipment can still safely accept in the current power allocation cycle.

[0118] S4.3. Based on the shrinkage coefficient corresponding to the status profile value, execution deviation, and neighborhood risk value, the remaining available power capacity is reduced. For devices with no remaining available power capacity, they will not participate in the subsequent capacity allocation, and a remaining capacity table will be generated.

[0119] Furthermore, based on the remaining capacity records of controllable devices, the device identifier, power boundary validity, status profile value, execution deviation value, neighborhood risk value, and remaining available power capacity of each device are determined. Power boundary validity is used as the basic condition for capacity retention, while status profile value, execution deviation value, and neighborhood risk value are used as the basis for capacity reduction. When the power boundary is invalid, the corresponding device is directly marked as having an invalid boundary and its available capacity is reduced. When the status profile value is lower than the preset status confidence threshold, it is determined to be a low-confidence status profile. When the execution deviation value is greater than the preset execution deviation threshold, it is determined to be an excessive execution deviation. When the neighborhood risk value is greater than the preset neighborhood risk threshold, it is determined to be an excessively high neighborhood risk, and the remaining available power capacity is reduced according to the corresponding reduction coefficient.

[0120] After capacity reduction, the remaining capacity of the equipment is checked. If the equipment has no remaining available power capacity, it is marked as an equipment that cannot participate in capacity allocation and is removed from the subsequent capacity allocation targets. If it still has remaining available power capacity, it continues to be an equipment that can participate in capacity allocation. For each equipment, the equipment identifier, source of capacity reduction, boundary validity, status profile value, execution deviation value, neighborhood risk value, remaining available power capacity, and allocation participation status are associated and saved to generate a remaining capacity table.

[0121] It should be noted that the state confidence threshold, execution deviation threshold, and neighborhood risk threshold can be set based on the equipment's historical normal operating conditions, scheduling allowable error, acquisition accuracy, and neighborhood power fluctuations.

[0122] A status profile value below 0.6 to 0.8 can be considered as low confidence in the status profile; an execution deviation value exceeding 3% to 8% of the equipment's rated active power can be considered as excessive execution deviation; and a neighborhood risk value above 0.6 to 0.8 can be considered as excessive neighborhood risk.

[0123] S4.4 Read the net target active power to be allocated, the remaining capacity table, the power boundary of controllable equipment, the allocated power, the status profile value and the equipment electrical-communication interaction diagram, and establish an iterative allocation ledger.

[0124] Furthermore, based on the net target power record to be allocated, the net target active power to be allocated is read. Based on the remaining capacity table, the equipment identifier, remaining available power capacity, and allocation participation status of each controllable device are read. Simultaneously, the power boundary, allocated power, status profile value, and connection relationship, adjacent device, and communication association status of the controllable device in the electrical-communication interaction diagram are read. The net target active power to be allocated is used as the total allocation benchmark for this round, the remaining available power capacity is used as the range that the device can accept, the power boundary and allocated power are used as the allocation constraints for a single device, and the status profile value and electrical-communication interaction relationship are used as the basis for iteration order and risk verification.

[0125] If a device is marked as not participating in allocation in the remaining capacity table, it will not be written into the valid iteration objects of this round; only the reason for removal will be retained. If the device still has remaining available power capacity, it will be included in the iterative allocation ledger, and the currently allocated power, additional capacity, upper and lower power boundaries, status profile value, and neighborhood relationship in the device's electrical-communication interaction diagram will be recorded. When there are anomalies in the device boundary, capacity status, or neighborhood communication relationship, the corresponding constraint source will be marked in the ledger. For devices that have completed registration, the device identifier, the source of net target active power to be allocated, the remaining capacity value, power boundary, allocated power, status profile value, interaction diagram association information, participation status, and constraint mark will be retained to establish the iterative allocation ledger.

[0126] S4.5. Generate initial allocation shares based on the remaining capacity ratio according to the iterative allocation ledger, and suppress the share of low-confidence devices to obtain candidate power allocation amounts.

[0127] It should be noted that the iterative allocation ledger is an allocation process record organized by device identifier, which records the capacity that can be allocated to each controllable device in the current round, the existing power occupancy, the allowable power range, and the operating status.

[0128] The remaining capacity percentage is the proportion of the remaining available power capacity of a single device in the total remaining available power capacity of all devices participating in the capacity percentage allocation.

[0129] Furthermore, based on the iterative allocation ledger, the net target active power to be allocated, the device identifier of each participating device, the remaining available power capacity, the power allocated in the previous cycle, the power boundary, the status profile value, the execution deviation value, and the neighborhood consistency result are determined. When the remaining available power capacity of a device is valid and the participation status is normal, an initial allocation share is generated according to the ratio of the remaining available power capacity of that device to the total remaining available power capacity of all participating devices. If the status profile value of a device is lower than the confidence requirement, it is marked as a low-confidence device and the initial allocation share is reduced. If the execution deviation value of a device exceeds the allowable range, it is marked as a high-deviation device and the allocation share is further suppressed. If the neighborhood consistency of a device is low, the allocation share is suppressed according to the degree of neighborhood inconsistency. If the neighborhood consistency is high, the allocation capacity of that device is retained or increased.

[0130] After the share suppression is completed, non-negative constraints and boundary checks are performed on the equipment. If the suppressed allocation amount is less than zero, it is set to zero. If the suppressed allocation amount exceeds the remaining available power capacity of the equipment or causes the target power of the equipment to exceed the power boundary, it is limited according to the remaining available power capacity or power boundary. If the equipment no longer has an effective capacity after suppression and limiting, it is marked as an invalid candidate allocation and excluded from the candidate allocation. For each equipment, the equipment identifier, remaining available power capacity, capacity ratio source, initial allocation share, status profile value, execution deviation value, neighborhood consistency result, share suppression reason, boundary limiting result and candidate power allocation amount are associated and saved to form a candidate power allocation record.

[0131] The calculation expression for the candidate power allocation is as follows:

[0132] ;

[0133] in, Indicates device Candidate power allocation; This refers to the set of controllable devices that participate in the allocation of capacity proportions. Represents a set Any device identifier in the list; Indicates device The remaining available power capacity; This represents the net target active power to be allocated; Indicates device The remaining available power capacity; Indicates device The low confidence suppression level; Indicates device Execution deviation; Indicates device The neighborhood consistency value; This represents the low-confidence suppression weight coefficient, which is set according to the degree of influence of the state profile value on the allocatable reliability of the device. This represents the execution deviation suppression weighting coefficient, which is set according to the degree of influence of the execution deviation between the actual power and the allocated power on the allocation result. This represents the neighborhood consistency correction weight coefficient, which sets the level of support for the power that the device can continue to receive based on neighborhood consistency.

[0134] It should be noted that the candidate power allocation amount represents the amount of power adjustment that the equipment can accept in the current adjustment direction, and is not less than zero by default. When the equipment supports bidirectional adjustment of charging / discharging or increasing / decreasing power, the positive and negative directions of the candidate power allocation amount are determined by the adjustment direction of the net target power to be allocated in the current round, and are limited according to the power boundary and remaining available power capacity of the corresponding direction.

[0135] S4.6 In each round, write the candidate power allocation amount back to the iterative allocation ledger, lock the edge devices, backfill the unallocated residuals, and generate the target active power after meeting the convergence conditions. Remove the devices that have reached the upper and lower power limits from the subsequent allocation set, and redistribute the unallocated residuals among the unlocked devices with remaining capacity according to the proportion of remaining capacity, until the residuals are less than the preset convergence threshold.

[0136] Furthermore, based on the candidate power allocation records, the device identifier, the power allocated in the previous cycle, the candidate power allocation amount, the power boundary, the remaining available power capacity, and the share suppression result of each device are determined. The power allocated in the previous cycle is superimposed with the candidate power allocation amount to obtain the current target power of the device, and the current target power of the device is written back to the iterative allocation ledger. If the current target power of the device reaches the power upper limit or the power lower limit, the device is marked as a locked device and will not be allocated in subsequent rounds. If the device is not locked and still retains remaining available capacity, it will continue to be a device that can participate in the allocation in the next round.

[0137] The power portion that could not be taken in this round due to edge locking, share suppression, or insufficient capacity is marked as unallocated residual, and the unallocated residual is fed back to the iterative allocation ledger. The equipment that has not been touched and still has remaining capacity continues to take over according to the proportion of remaining capacity. If the unallocated residual is less than the preset convergence threshold, or there is no equipment in the iterative allocation ledger that can continue to take over, the iterative allocation ends. The final result is associated with and saved with the equipment identifier, the candidate power allocation amount in each round, the current target power, the edge status, the source of the residual, the feedback status, the convergence reason, and the final target active power, and a target active power record is generated.

[0138] It should be noted that the convergence condition can be preset based on the power scheduling accuracy, the minimum adjustable step size of the equipment, the residual allowable error and the maximum number of iterations. Usually, the termination condition is that the unallocated residual is less than the minimum adjustable power or the maximum number of iterations is reached.

[0139] S5. Combine the target active power of each controllable device with the fixed power of each uncontrollable device to generate device-level power control commands.

[0140] S5.1. Read the target power of controllable devices and the fixed power of uncontrollable devices from the iterative allocation ledger and the uncontrollable device fixed power list, and establish a device-level instruction ledger.

[0141] Furthermore, the system retrieves the device identifier, target active power, and allocation status of controllable devices from the iterative allocation ledger, and the device identifier, fixed power, and fixed reason of uncontrollable devices from the fixed power list of uncontrollable devices. The target active power is used as the basis for adjustment instructions for controllable devices, and the fixed power is used as the basis for power locking of uncontrollable devices. If the device is controllable, it is registered as an adjustable instruction; if the device is uncontrollable, it is registered as a fixed record.

[0142] If the same device appears in both the controllable device record and the uncontrollable device record, it is determined to be a cluster consistency anomaly. The issuance of adjustment commands to that device is stopped, and the device is transferred to the uncontrollable device set. The fixed power of the uncontrollable device is used as the basis for power processing of that device in this round, and the reason for the command conflict is marked. The processing results are associated and saved with the device identifier, device source, target power or fixed power, processing reason, conflict mark, and final command power, establishing a device-level command ledger.

[0143] S5.2 Perform boundary verification, step size shaping, and instruction type marking on the merged power, and encapsulate power dispatch instructions based on the instruction ledger to generate device-level power control instructions.

[0144] Furthermore, based on the device-level instruction ledger, the device identifier, target power, power boundary, adjustment step size, and control attributes of each controllable device are obtained. The power boundary is used as the verification basis, and the adjustment step size is used as the shaping basis. If the target power exceeds the power boundary, it is limited and marked as boundary correction. If the target power does not meet the adjustment step size requirements, it is shaped and marked as step size correction. Controllable devices that have completed boundary verification and step size shaping are marked as adjustment instructions, and device-level power control instructions are generated.

[0145] For uncontrollable equipment, no regulating power control command is generated. Instead, a fixed power record is retained in the equipment-level command ledger or a monitoring record is generated to characterize the power lock-in status of the equipment in the current allocation cycle. The processing results are associated with and saved as follows: equipment identifier, power before correction, power after correction, reason for correction, command type, fixed power record, and final issued power.

[0146] In summary, this invention improves scheduling accuracy by using an electrical-communication interaction diagram and deviation closed-loop correction to enable microgrid clusters under distributed control to dynamically adjust the total target power based on actual response, communication reliability, and electrical adjacency. Furthermore, by dividing the controllable and uncontrollable device sets using the current-round allocable latch value, and combining fixed power deduction, capacity contraction, and iterative allocation, it avoids power allocation falling into uncontrollable or low-reliability devices, thus enhancing the security and executability of device-level power control commands.

[0147] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention, and all such modifications or substitutions should be covered within the scope of the claims of the present invention.

Claims

1. A microgrid cluster coordinated power allocation method based on closed-loop feedback, characterized in that, include: Collect status data of microgrid equipment clusters, standardize and encapsulate the status data, verify its validity and construct status profiles to generate an equipment status matrix. The equipment status matrix includes equipment identification, online status, controllable status, actual active power, power boundary parameters and active power allocated in the previous cycle. Based on the actual total response active power in the equipment state matrix, the deviation between the actual total response active power and the theoretical total target active power is calculated, and the theoretical total target active power is corrected in a closed loop to obtain the total target active power corresponding to the current power allocation cycle. Based on the online and controllable status in the equipment status matrix, offline equipment is excluded from this round of allocation. Online and controllable equipment is divided into a controllable equipment set, and online but uncontrollable equipment is divided into an uncontrollable equipment set. The current actual operating power of each equipment in the uncontrollable equipment set is used as the fixed power. The fixed power is deducted from the corrected total target active power to obtain the net target active power to be allocated. Extract power boundary parameters and allocated power of the controllable device set from the device state matrix, calculate the remaining available power capacity of each controllable device, and perform iterative allocation of the net target active power to be allocated with boundary constraints according to the proportion of the remaining available power capacity of each controllable device to the total remaining available power capacity, to obtain the target active power. The target active power of each controllable device is combined with the fixed power of each uncontrollable device to generate device-level power control commands.

2. The microgrid cluster coordinated power allocation method based on closed-loop feedback as described in claim 1, characterized in that, The process involves collecting status data from the microgrid device cluster, standardizing and encapsulating the status data, verifying its validity, and constructing a status profile to generate a device status matrix. This matrix includes device identifier, online status, controllable status, actual active power, power boundary parameters, and the active power allocated in the previous cycle. The specific steps are as follows: Status data is collected from the equipment sample library and then time-aligned, unit-unified, and field-encapsulated to obtain standardized status records. The standardized status records are validated and status profiles are constructed. The status profile values ​​are generated based on online status, communication quality, command response residuals, and power boundary validity. The online status, controllable status, power boundary, and actual power are integrated according to the device identifier to generate a device status matrix.

3. The microgrid cluster coordinated power allocation method based on closed-loop feedback as described in claim 2, characterized in that, The deviation between the actual total response active power and the theoretical total target active power is calculated based on the actual total response active power in the device state matrix. The specific steps are as follows: Power, status, communication, and boundary data are read from the device status matrix and device sample library, and the device response sequence is obtained by filtering. Construct an electrical-communication interaction diagram for the equipment, perform consistency correction and residual filtering on the equipment response sequence, and generate deviation values, deviation levels, and abnormal equipment identifiers.

4. The microgrid cluster coordinated power allocation method based on closed-loop feedback as described in claim 3, characterized in that, The specific steps for obtaining the total target active power corresponding to the current power allocation cycle are as follows: Based on the deviation value, deviation level, and abnormal equipment identifier, combined with the total target active power corrected in the previous cycle, the current theoretical total target power, and the power carrying capacity boundary, a closed-loop correction trigger quantity is generated. The total target power correction constraint domain is formed based on the closed-loop correction trigger value. Within the total target power correction constraint domain, the theoretical total target power is recursively corrected based on the historical deviation trend, the continuity of the correction direction, and the equipment reliability status to obtain the total target active power corresponding to the current power allocation cycle.

5. The microgrid cluster coordinated power allocation method based on closed-loop feedback as described in claim 3, characterized in that, The device electrical-communication interaction diagram is constructed by using each device as a node, establishing electrical connection edges based on the bus assignment and feeder connection relationships in the device sample library, determining communication reliability based on heartbeat, communication latency and packet loss rate in the network status, and then adjusting the node weights.

6. The microgrid cluster coordinated power allocation method based on closed-loop feedback as described in claim 5, characterized in that, Based on the online and controllable states in the device status matrix, offline devices are excluded from this round of allocation. Online and controllable devices are divided into a controllable device set, and online but uncontrollable devices are divided into an uncontrollable device set. The current actual operating power of each device in the uncontrollable device set is used as the fixed power. The specific steps are as follows: State feature data is read from the device state matrix, and adjacency relationships and edge reliability coefficients are extracted by combining the device electrical-communication interaction graph to obtain the basic data for device clustering. Based on the basic data of device clustering, basic state trust features, communication trust features, residual trust features, neighborhood consistency features and power boundary security features are generated to form a core feature group. The core feature group is then weighted and normalized to obtain the allocatable latch value for this round. If a device is offline, it will be excluded from this round of allocation. If a device is online and controllable, and the allocatable latch value in this round reaches the latch threshold, it is classified as a controllable device set. If a device is online but uncontrollable, or if a device is online and controllable but the available latch value for this round has not reached the latch threshold, it is classified as an uncontrollable device set. The current actual active power of each device in the set of uncontrollable devices is locked as the fixed power, and the device identifier, fixed power value and fixed reason are stored accordingly.

7. The microgrid cluster coordinated power allocation method based on closed-loop feedback as described in claim 6, characterized in that, The specific steps for subtracting the fixed power from the corrected total target active power to obtain the net target active power to be allocated are as follows: Read the corrected total target power and the uncontrollable equipment curing power, and perform risk equivalence processing on the curing power to obtain the equivalent curing power; If the set of controllable devices is empty, the power allocation of controllable devices in this round is stopped, and the fixed power of uncontrollable devices is used as the output power command for this round. If the set of controllable devices is not empty, the equivalent fixed power is deducted and the boundary is limited to obtain the net target power to be allocated.

8. The microgrid cluster coordinated power allocation method based on closed-loop feedback as described in claim 7, characterized in that, The steps for extracting power boundary parameters and allocated power of the controllable device set from the device state matrix, and calculating the remaining available power capacity of each controllable device are as follows: Extract the power boundary, allocated power, actual power, and status profile values ​​of controllable devices from the device status matrix to generate a capacity ledger; By combining boundary margin, ramp achievable amount, execution deviation and neighborhood consistency, the remaining available power capacity of each controllable device is calculated; The remaining available power capacity is reduced according to the shrinkage coefficient corresponding to the status profile value, execution deviation, and neighborhood risk value. For devices with no remaining available power capacity, they will not participate in the subsequent capacity allocation, and a remaining capacity table will be generated.

9. The microgrid cluster coordinated power allocation method based on closed-loop feedback as described in claim 8, characterized in that, The specific steps to obtain the target active power are as follows: Read the net target active power to be allocated, the remaining capacity table, the power boundary of controllable equipment, the allocated power, the status profile value, and the equipment electrical-communication interaction diagram, and establish an iterative allocation ledger; The initial allocation share is generated based on the percentage of remaining capacity according to the iterative allocation ledger, and the share is suppressed for low-confidence devices to obtain the candidate power allocation amount. In each round, the candidate power allocation is written back to the iterative allocation ledger, the edge devices are locked, the unallocated residuals are fed back, and the target active power is generated after the convergence condition is met. Devices that have reached the upper and lower power limits are removed from the subsequent allocation set, and the unallocated residuals are redistributed among the unlocked devices with remaining capacity according to the proportion of remaining capacity, until the residuals are less than the preset convergence threshold.

10. The microgrid cluster coordinated power allocation method based on closed-loop feedback as described in claim 9, characterized in that, The specific steps for generating device-level power control commands are as follows: Read the target power of controllable devices and the fixed power of uncontrollable devices from the iterative allocation ledger and the list of fixed power of uncontrollable devices, and establish a device-level instruction ledger; The merged power is subjected to boundary verification, step size shaping and instruction type marking, and power dispatch instructions are encapsulated based on the instruction ledger to generate device-level power control instructions.

Images