A cloud-edge collaborative control method for power distribution load energy efficiency optimization

Abstract

The application discloses a kind of power distribution load energy efficiency optimization's cloud edge collaborative control method, it is related to electrical equipment state monitoring technical field, the present application is by the feeder load state, cabinet thermal constraint state and load switchable state are written in the basic graph atlas of cabinet group load partial load graph, and form the associated mark containing state source, cloud control platform can simultaneously according to overload association, temperature rise association, light load association and forbidden investment association when generating progressive load balancing sequence determine the object to be released and can be carried object;Edge control node is again with edge security envelope screening action, and action execution result and state summary after execution are written back to cabinet group load partial load graph, so that subsequent progressive load balancing sequence is generated according to the cabinet group state after execution, to make feeder partial load processing, cabinet thermal constraint avoidance and switching result write formation continuous control chain.

Description

Technical Field

[0001] This invention relates to the field of electrical equipment condition monitoring technology, and in particular to a cloud-edge collaborative control method for optimizing the energy efficiency of power distribution loads. Background Technology

[0002] In older power distribution rooms, multi-tenant commercial buildings, and existing industrial parks, tenant business types and electricity usage periods change frequently. The connection time, duration, and power fluctuations of loads such as catering, air conditioning, cold storage, lighting, and charging equipment vary greatly, easily leading to a situation where different feeders in the same power distribution room experience both heavy and light loads. With the application of intelligent switchgear and cloud-edge collaborative platforms, the edge side can collect cabinet-side operating data such as current, voltage, power, switch status, cabinet temperature rise, alarm status, and communication status. The cloud can aggregate data from multiple cabinets and generate load scheduling strategies.

[0003] Existing power distribution load control methods typically rely on feeder power, load level, or historical electricity consumption curves as primary data sources. Control strategies are generated in the cloud, and then the edge devices execute connection, disconnection, or restoration actions according to fixed thresholds, preset priorities, or single control commands. This type of method can achieve basic load monitoring and localized adjustments when the feeder load distribution is relatively stable.

[0004] Under conditions where feeder off-center load and cabinet temperature rise coexist, existing methods lack unified constraints on feeder off-center load relationships, cabinet thermal status, load switchability attributes, and real-time execution status at the edge. Switchgear at risk of temperature rise, with alarms not yet cleared, or in a cooling-off state may still be included in the access or recovery candidate list. After the edge executes an action, there is no clear state write-back relationship between the cloud strategy and the actual action result. This means that subsequent scheduling may still generate new switching actions based on the already failed cabinet group status, leading to repeated load migration, premature triggering of recovery actions, and continuous accumulation of switchgear thermal risks.

[0005] Therefore, it is necessary to address the lack of a closed-loop connection between cloud load balancing strategies, edge security screening, and write-back of switching results when both feeder off-center load and cabinet thermal constraints coexist. Summary of the Invention

[0006] This application provides a cloud-edge collaborative control method for optimizing power distribution load energy efficiency, which solves the problem of unclear cloud-edge switching closed loop under feeder off-center load and cabinet thermal constraints.

[0007] This invention provides a cloud-edge collaborative control method for optimizing power distribution load energy efficiency, applicable to a power distribution system including a cloud control platform, edge control nodes, and intelligent switchgear, comprising:

[0008] The edge control node collects cabinet-end operation data and generates cabinet load records and cabinet-level energy efficiency off-load status.

[0009] The cloud-based control platform constructs a load eccentricity map of the cabinet group based on the cabinet load records and the cabinet-level energy efficiency load eccentricity status. The load eccentricity map of the cabinet group includes switch cabinet nodes, feeder nodes, load unit nodes, and basic map edges. The cloud-based control platform writes the feeder load status, cabinet thermal constraint status, and load switchability status into the basic map edges to form at least one of the following association markers containing the status source: heavy load association, light load association, temperature rise association, and no-operation association.

[0010] When there are objects to be released marked by at least one of heavy load association and temperature rise association in the load distribution diagram of the cabinet group, and there are load-bearing objects marked by light load association but not marked by prohibited loading association, the cloud control platform generates a progressive load balancing sequence based on the objects to be released, load-bearing objects and the thermal constraint status of the cabinet.

[0011] The edge control node forms an edge safety envelope based on the cabinet temperature rise status, alarm status, communication status, switch action cooling status, and load switchability status. It then filters the actions in the progressive load balancing sequence, executes the filtered actions, and uploads the action execution results and post-execution status summary.

[0012] The cloud-based control platform updates the load eccentricity map of the cabinet group based on the action execution results and the post-execution status summary, and uses the updated load eccentricity map of the cabinet group as the basis for generating the subsequent progressive load balancing sequence.

[0013] In some embodiments, cabinet operation data includes current, voltage, power, switch status, cabinet temperature rise status, alarm status, and communication status; cabinet load records include switch cabinet identifier, feeder affiliation, load unit identifier, power user identifier, phase identifier, acquisition time, and load attributes.

[0014] A load unit is an electrical load or a collection of electrical loads that is connected to the intelligent switch cabinet feeder and can be identified as an object to be connected, disconnected, or restored.

[0015] Edge control nodes collect cabinet operation data within the same acquisition cycle according to switch cabinet identification, feeder affiliation, and load unit identification.

[0016] In some embodiments, load attributes include at least one of guaranteed load attributes, interruptible load attributes, delayed access load attributes, recoverable load attributes, and restricted switching load attributes;

[0017] The edge control node writes the load attributes into the corresponding load unit identifier and saves the cabinet load record in the local recording area; the local recording area saves the record version of the cabinet load record according to the switch cabinet identifier and the time of acquisition;

[0018] When communication between the edge control node and the cloud control platform fails, the edge control node verifies the received progressive load balancing sequence using the cabinet load record in its local recording area.

[0019] In some embodiments, when the edge control node generates a cabinet-level energy efficiency off-load state, it determines the feeder load state according to the feeder load range, determines the phase balance state according to the load difference of each phase, determines the cabinet thermal constraint state according to the cabinet temperature rise state and alarm state, determines the switch action cooling state according to the holding time after the most recent switch action, and determines the load switchable state according to the load attributes.

[0020] The cabinet's thermal constraint status includes accessible status, restricted access status, prohibited operation status, and disconnection-only status; the switch operation cooling status includes cooling hold status and operation permission status.

[0021] In some embodiments, the load distribution diagram of the cabinet group also includes the main power consumption node and the phase node;

[0022] The basic diagram edges include the power supply attribution edge between switch cabinet nodes and feeder nodes, the access attribution edge between feeder nodes and load unit nodes, the power consumption attribution edge between load unit nodes and power-consuming entity nodes, the phase attribution edge between load unit nodes and phase nodes, the same cabinet relationship edge, the same feeder relationship edge, the same power-consuming entity relationship edge, and the same phase relationship edge.

[0023] The cloud control platform writes the feeder load status into the basic graph edge related to the feeder node, writes the cabinet thermal constraint status into the basic graph edge related to the switch cabinet node, and writes the load switchability status into the basic graph edge related to the load unit node.

[0024] In some embodiments, when the cloud control platform generates a progressive load balancing sequence, it forms a set of objects to be released and a set of objects that can be carried based on the load eccentricity map of the cabinet group used when the sequence is generated, and reads the recovery candidate set that matches the map version identifier of the load eccentricity map of the cabinet group.

[0025] The set of objects to be released includes switch cabinet identifier, feeder affiliation, association type, status source, and load unit range. The association type includes heavy load association and temperature rise association.

[0026] The set of load-bearing objects includes switchgear identification, feeder affiliation, light load association, prohibited operation association mark, and action permission status. The set of recovery candidates includes load unit identification, target switchgear, target feeder, recovery source, most recent disconnection action, and cabinet thermal constraint status.

[0027] In some embodiments, the cloud control platform generates a progressive load balancing sequence in the order of restricting new access, releasing interruptible loads, and restoring recoverable loads;

[0028] The cloud control platform first processes new access requests based on the set of objects to be released, and generates delayed access actions or disallow access actions;

[0029] When the heavy load association or temperature rise association corresponding to the object to be released still exists, the cloud control platform generates a disconnection action based on the load switchable status.

[0030] When the cloud control platform determines that the overload association and temperature rise association corresponding to the object to be released have been released based on the action execution result and the post-execution status summary, the cloud control platform updates the object to be released to the released status and generates a recovery action based on the recovery candidate set;

[0031] Each action in the progressive load balancing sequence includes the target switchgear, target feeder, load unit, action type, action sequence, and triggering basis.

[0032] In some embodiments, when the edge control node performs screening, it rejects new access actions if the target switchgear has abnormal temperature rise or the alarm has not been cleared; it rejects actions again if the target switchgear or target feeder is in a switching operation cooling state; it rejects disconnection actions if the target load unit has load supply protection attributes or load switching restriction attributes; it rejects target actions if the target action causes the phase-to-phase balance state to become unfavorable; and it rejects access actions, disconnection actions, and recovery actions if the target switchgear communication state is abnormal and the action result cannot be confirmed. The edge control node generates rejection reasons for the rejected target actions.

[0033] In some embodiments, when the edge control node performs the filtered action, it performs access, disconnection or recovery control item by item in the order of the progressive load balancing sequence, and reads the switch status, cabinet temperature rise status, alarm status, feeder load status and phase balancing status of the target switch cabinet after each action is performed, and generates a post-execution status summary based on the reading results.

[0034] When the execution result of the action is inconsistent with the target action, the edge control node stops executing subsequent actions of the same target switch cabinet or the same target feeder and generates an abnormal execution reason; when the thermal constraint status of the cabinet after execution is an accessible state and the target load unit has a recoverable load attribute, the edge control node adds the target load unit to the recovery candidate set.

[0035] In some embodiments, the cloud control platform sets a map version identifier for the load off-center map of the cabinet group; the action execution results uploaded by the edge control node include the sequence version identifier, target switch cabinet, target feeder, load unit, action type, execution status and execution time;

[0036] The execution status includes one of the following: executed, rejected, abnormal execution, and suspended.

[0037] The rejected status corresponds to the reason for rejection, and the abnormal execution status corresponds to the reason for abnormal execution.

[0038] The post-execution status summary includes the post-execution feeder load status, post-execution phase-to-phase balance status, post-execution cabinet temperature rise status, post-execution alarm status, post-execution cabinet thermal constraint status, and post-execution switch action cooling status.

[0039] The cloud control platform writes the action execution results and post-execution status summary into the cabinet group load eccentricity map with the corresponding map version identifier, and uses the target actions corresponding to the rejected state and abnormal execution state as the basis for the cloud control platform to select target actions when generating subsequent progressive load balancing sequences.

[0040] Through the above technical solution, the present invention can achieve at least the following beneficial effects:

[0041] This invention writes the feeder load status, cabinet thermal constraint status, and load switchability status into the basic graph edge of the cabinet group load eccentricity graph, forming an association marker containing the status source. When generating a progressive load balancing sequence, the cloud control platform can simultaneously determine the objects to be released and the objects that can be carried based on heavy load association, temperature rise association, light load association, and prohibited operation association. The edge control node then filters actions with edge safety envelopes and writes the action execution results and post-execution status summary back to the cabinet group load eccentricity graph, so that the subsequent progressive load balancing sequence is generated based on the cabinet group status after execution. This makes the feeder load eccentricity processing, cabinet thermal constraint avoidance, and operation switching result writeback form a continuous control chain.

[0042] After the cabinet load record is written with the switch cabinet identifier, feeder affiliation, load unit identifier, acquisition time, and load attributes, the edge control node can aggregate the cabinet operation data to the determined load unit and feeder object. In case of communication failure, the record version in the local recording area can be used to verify the received progressive load balancing sequence, so that the action execution basis corresponds to the local cabinet status.

[0043] After the cabinet-level energy efficiency off-load status includes feeder load status, phase-to-phase balance status, cabinet thermal constraint status, switch action cooling status, and load switchable status, the edge control node can identify abnormal temperature rise, unresolved alarms, cooling maintenance, and load attribute conflicts before action screening, so that new access, disconnection, and recovery actions are constrained by the cabinet-side status before execution.

[0044] After the progressive load balancing sequence is generated in the order of restricting new access, releasing interruptible loads, and restoring recoverable loads, the cloud control platform first processes the new access requests on the objects to be released, then generates a disconnection action when the off-load association still exists, and generates a recovery action after the overload association and temperature rise association are removed, so that the triggering of the recovery action depends on the execution of the post-state summary.

[0045] After the action execution result, rejection reason, abnormal execution reason, and post-execution status summary are written into the local execution record along with the sequence version identifier and uploaded, the cloud control platform can use the target actions corresponding to the rejected state and abnormal execution state as the basis for the cloud control platform to select target actions when generating subsequent progressive load balancing sequences, so that action objects that have been rejected by the edge security envelope or have executed abnormally are included in the state constraints generated by the subsequent sequence. Attached Figure Description

[0046] To more clearly illustrate the technical solutions of the embodiments of this application, the accompanying drawings used in the embodiments will be briefly described below. It should be understood that the following drawings only show some embodiments of this application and should not be regarded as a limitation on the scope of this application.

[0047] Figure 1 This is a flowchart of the cloud-edge collaborative control method for optimizing power distribution load energy efficiency in the embodiments. Detailed Implementation

[0048] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description is provided in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the scope of this application.

[0049] All terms used in this application (including technical and scientific terms) have the meanings commonly understood by those skilled in the art, unless otherwise defined. It should be noted that the terms used herein should be interpreted in a manner consistent with the context of this specification, and not in an idealized or overly rigid way.

[0050] To facilitate understanding, the relevant terms and concepts involved in the embodiments of this application will be introduced below:

[0051] Cabinet-side operation data refers to the operation data collected by the intelligent switch cabinet or read from the intelligent switch cabinet by the edge control node. The data categories include electrical quantities, switch status, cabinet temperature rise status, alarm status, and communication status.

[0052] Cabinet load records refer to data records collected by edge control nodes according to the same switchgear, feeder, and load unit. Fields include switchgear identifier, feeder affiliation, load unit identifier, power user identifier, phase identifier, collection time, and load attributes. A load unit refers to a power load or set of power loads connected to the intelligent switchgear feeder and identifiable as an object to be connected, disconnected, or restored.

[0053] Cabinet-level energy efficiency off-load status refers to the set of cabinet-side statuses generated by the edge control node based on cabinet-side operating data. This set includes feeder load status, phase-to-phase balance status, cabinet thermal constraint status, switch operation cooling status, and load switchability status. Cabinet group load off-load map refers to the cabinet group relationship data structure maintained by the cloud control platform.

[0054] A progressive load balancing sequence refers to an action sequence generated by the cloud control platform based on the associated markers in the cabinet group load distribution map, the cabinet thermal constraint status, and the load switchability status. An edge safety envelope refers to the set of state constraints adopted by the edge control node before action execution, which includes cabinet temperature rise status, alarm status, communication status, switch action cooling status, and load switchability status.

[0055] Furthermore, the record version refers to the data version identifier written by the edge control node when generating cabinet load records and cabinet-level energy efficiency off-load status. The same record version corresponds to the cabinet-side operating data within the same collection period. The graph version identifier refers to the data version identifier written by the cloud control platform when updating the cabinet group load off-load graph. The same graph version identifier corresponds to the node status and basic graph edge status after a graph update. The sequence version identifier refers to the data version identifier written by the cloud control platform when generating a progressive load balancing sequence. The same sequence version identifier corresponds to the action sequence, triggering basis, and graph version identifier used during generation.

[0056] The electricity user entity refers to the tenant, management area, and electricity collection object jointly determined by the tenant and management area corresponding to the load unit. The electricity user entity identifier is a data field used to identify the electricity user entity. The electricity user entity node is the node in the cabinet group load off-center load diagram that corresponds to the electricity user entity. The relationship with the same electricity user entity is the basic graph edge type established when two load unit nodes correspond to the same electricity user entity.

[0057] Example 1:

[0058] like Figure 1 As shown, this embodiment employs a cloud-edge collaborative control method for optimizing power distribution load energy efficiency. It is applied to a power distribution system including a cloud control platform, edge control nodes, and multi-faceted intelligent switchgear. The edge control nodes communicate with the intelligent switchgear within their control domain. The method includes:

[0059] Step S1: The edge control node collects the cabinet-end operation data of the intelligent switchgear in the control domain, generates cabinet load records, and generates cabinet-level energy efficiency off-load status based on the cabinet-end operation data.

[0060] In step S2, the cloud control platform constructs a load eccentricity map of the cabinet group based on the cabinet load records and the cabinet-level energy efficiency load eccentricity status. The load eccentricity map of the cabinet group includes switch cabinet nodes, feeder nodes, load unit nodes, main power consumption nodes, phase nodes, and basic map edges. The cloud control platform writes the feeder load status, cabinet thermal constraint status, and load switchable status into the corresponding basic map edges, forming at least one association marker among the following: heavy load association, light load association, temperature rise association, connectable association, disconnectable association, and prohibited operation association, which contains the status source.

[0061] Step S3: When there are objects to be released marked by at least one of heavy load association and temperature rise association in the load distribution map of the cabinet group, and there are load-bearing objects marked by light load association but not marked by prohibited loading association, the cloud control platform generates a progressive load balancing sequence based on the objects to be released, load-bearing objects and the corresponding cabinet thermal constraint status.

[0062] Step S4: The edge control node filters the actions in the progressive load balancing sequence based on the edge safety envelope formed by the cabinet temperature rise status, alarm status, communication status, switch action cooling status and load switchability status, executes the actions that pass the filter, and uploads the action execution results and post-execution status summary to the cloud control platform.

[0063] In step S5, the cloud control platform updates the load eccentricity map of the cabinet group based on the action execution results and the post-execution status summary, and uses the updated load eccentricity map of the cabinet group as the basis for generating the subsequent progressive load balancing sequence.

[0064] In one implementation, when the cloud control platform constructs the load distribution map of the switchgear group, it records the connection relationship between switchgear nodes and feeder nodes as power supply ownership edge, the connection relationship between feeder nodes or load unit nodes within the same intelligent switchgear as same-cabinet relationship edge, the connection relationship between load unit nodes under the same feeder as same-feeder relationship edge, the connection relationship between load unit nodes and power-consuming entity nodes as power consumption ownership edge, the connection relationship between different load unit nodes under the same power-consuming entity as same-power-consuming entity relationship edge, the connection relationship between load unit nodes and phase nodes as phase ownership edge, and the connection relationship between different load unit nodes under the same phase as same-phase relationship edge. After receiving the cabinet-level energy efficiency off-load status, the cloud control platform writes the feeder load status into the basic graph edge related to the feeder node, writes the cabinet thermal constraint status into the basic graph edge related to the switch cabinet node, and writes the load switchability status into the basic graph edge related to the load unit node. It also records the edge control node identifier, record version, acquisition time, and status source on the corresponding node or graph edge. The status source refers to the cabinet load record on which the corresponding node status or graph edge status is based.

[0065] In one implementation, the cabinet operation data includes current, voltage, power, switch status, cabinet temperature rise status, alarm status, and communication status; the cabinet load record includes switch cabinet identification, feeder affiliation, load unit identification, power user identification, phase identification, acquisition time, and load attributes; the load unit is a power load or set of power loads that is connected to the intelligent switch cabinet feeder and can be identified as an access, disconnection, or restoration object; the edge control node collects the cabinet operation data within the same acquisition period according to the switch cabinet identification, feeder affiliation, and load unit identification.

[0066] Specifically, the acquisition period refers to the time interval during which the edge control node collects data from the intelligent switchgear within the control domain. The acquisition period parameter ranges from 1 second to 60 seconds, and the edge control node writes the acquisition time using the same clock source. Current, voltage, and power are formed using valid sampled values ​​within the same acquisition period; if multiple valid sampled values ​​exist within the same acquisition period, the edge control node uses the average value within that acquisition period as the corresponding electrical quantity. When the communication status is abnormal or a field is missing, the edge control node retains a missing marker for the corresponding field and writes this missing marker into the cabinet load record.

[0067] In one implementation, the load attributes include at least one of the following: guaranteed load attributes, interruptible load attributes, delayed access load attributes, recoverable load attributes, and restricted switching load attributes; the edge control node writes the load attributes into the corresponding load unit identifier and saves the cabinet load record in the local recording area; the local recording area saves the record version of the cabinet load record according to the switch cabinet identifier and the acquisition time; when the edge control node communicates abnormally with the cloud control platform, the edge control node verifies the received progressive load balancing sequence with the cabinet load record in the local recording area.

[0068] When the edge control node generates the cabinet-level energy efficiency off-load status, it determines the feeder load status according to the feeder load range, the phase balance status according to the load difference between each phase, the cabinet thermal constraint status according to the cabinet temperature rise status and alarm status, the switch cooling status according to the holding time after the most recent switch action, and the load switchable status according to the load attributes. The load switchable status includes at least one of the following: allowed delayed access status, allowed disconnection status, allowed recovery status, prohibited disconnection status, and prohibited access status. The edge control node generates the load switchable status of the corresponding load unit according to the load attributes. Among them, the load attribute that can be delayed access corresponds to the allowed delayed access status, the load attribute that can be interrupted corresponds to the allowed disconnection status, the load attribute that can be recovered corresponds to the allowed recovery status, the load attribute that can be guaranteed to supply corresponds to the prohibited disconnection status, and the load attribute that restricts switching corresponds to at least one of the prohibited access status and prohibited disconnection status.

[0069] The cabinet's thermal constraint status includes accessible status, restricted access status, prohibited operation status, and disconnection-only status; the switch operation cooling status includes cooling hold status and operation permission status.

[0070] The switch action cooling state refers to the state result generated by the edge control node based on the holding time after the most recent access, disconnection, or recovery action of the target switch cabinet or feeder. The holding time is an implementation parameter, ranging from 30 seconds to 1800 seconds. When the holding time after the most recent switch action is less than the required holding time, the edge control node generates a cooling holding state; when the holding time after the most recent switch action reaches the required holding time, and both the switch status and communication status can be confirmed, the edge control node generates an action permission state.

[0071] Cabinet temperature rise status refers to the state result of the temperature acquisition value inside the intelligent switch cabinet relative to the corresponding temperature rise limit. Alarm status refers to the unresolved alarm result recorded by the intelligent switch cabinet or edge control node. The edge control node generates an accessible status when the cabinet temperature rise is below the temperature rise limit and the alarm status is resolved; a restricted access status when the cabinet temperature rise reaches the warning range and the alarm status is resolved; a prohibited operation status when the cabinet temperature rise reaches the prohibited operation range or there is an unresolved alarm related to switch operation; and a disconnection-only status when the cabinet temperature rise reaches the restricted access range and the disconnection action is used to reduce load. The temperature rise limit and alarm type use any one of the following as the data source: intelligent switch cabinet factory configuration, field protection settings, or operation and maintenance configuration.

[0072] Phase-to-phase equilibrium refers to the state resulting from the load differences between phases within the same switchgear or feeder. The edge control node calculates the maximum load difference between phases using the same data category for each phase's current or power, and compares this maximum load difference with the phase-to-phase difference threshold to generate an equilibrium or unfavorable state. The phase-to-phase difference threshold is an implementation parameter, ranging from 10% to 30% of the three-phase average load. When the maximum load difference between phases after the target action is executed exceeds the maximum load difference before the target action, and the increase reaches the phase-to-phase difference threshold, the edge control node determines that the target action has turned the phase-to-phase equilibrium state into an unfavorable state.

[0073] The feeder load range refers to the load status determination range defined based on the feeder's rated capacity and operating margin. The feeder operating margin is an implementation parameter, ranging from 5% to 20% of the feeder's rated capacity. The edge control node compares the feeder power or feeder current with the feeder load range to generate heavy load, light load, or normal load status. When both feeder power and feeder current are available, the status with the higher corresponding load level is used as the feeder load status. When one of them is missing, the status corresponding to the available data is used as the feeder load status and written into the status source.

[0074] In one implementation, when the edge control node performs screening, it rejects new access actions if the target switchgear has abnormal temperature rise or the alarm has not been cleared; it rejects actions again if the target switchgear or target feeder is in a switching operation cooling state; it rejects disconnection actions if the target load unit has load supply guarantee attributes or load switching restriction attributes; it rejects target actions if the target action changes the phase balance state to an unfavorable state; and it rejects access actions, disconnection actions, and recovery actions if the target switchgear communication state is abnormal and the action result cannot be confirmed.

[0075] The edge control node generates a rejection reason for the rejected target action.

[0076] When edge control nodes use edge safety envelopes to filter target actions, abnormal temperature rise, unresolved alarms, switch action cooling, load attribute conflicts, unfavorable changes in phase-to-phase balance, and inability to confirm action results are considered abnormal operating conditions. Rejection reasons include sequence version identifier, target switchgear, target feeder, load unit, action type, abnormal operating condition type, rejection time, and status source. After generating rejection reasons, the edge control node writes them to its local execution log and uploads them to the cloud control platform when communication with the cloud control platform is normal.

[0077] Specifically, abnormal temperature rise refers to the cabinet temperature rising reaching any of the following states: warning range, restricted access range, or prohibited operation range. Unresolved alarms refer to alarm records related to switchgear operation, temperature rise, communication, or contact status that have not been cleared. Load attribute conflict refers to a discrepancy between the target action and the permitted action range defined by the load attribute. Action result cannot be confirmed refers to the edge control node failing to read the switch status of the target switchgear after action execution, or reading a switch status inconsistent with the expected state corresponding to the target action.

[0078] In one implementation, when the edge control node performs the filtered action, it performs access, disconnection or recovery control item by item in the order of the progressive load balancing sequence, and reads the switch status, cabinet temperature rise status, alarm status, feeder load status and phase balance status of the target switch cabinet after each action is performed, and generates a post-execution status summary based on the reading results.

[0079] When the execution result of the action is inconsistent with the target action, the edge control node stops executing subsequent actions of the same target switch cabinet or the same target feeder and generates an abnormal execution reason; when the thermal constraint status of the cabinet after execution is an accessible state and the target load unit has a recoverable load attribute, the edge control node adds the target load unit to the recovery candidate set.

[0080] After the edge control node executes the target action, it reads the switch status, cabinet temperature rise status, alarm status, feeder load status, and phase balance status of the target switch cabinet, and compares the reading results with the expected status corresponding to the target action.

[0081] The expected state refers to the switchgear state that the target switchgear should exhibit after the target action is completed, as well as the corresponding load unit connection state. The expected state for connection and recovery actions is that the target load unit is in the connection state; the expected state for disconnection actions is that the target load unit is in the disconnection state; and the expected state for delayed connection and disabling actions is that no new target load unit has been added. The edge control node generates an executed state, a rejected state, an abnormal execution state, or a paused state based on the comparison between the expected state and the read result.

[0082] In one implementation, the action execution result uploaded by the edge control node includes a sequence version identifier, target switch cabinet, target feeder, load unit, action type, execution status, and execution time; when the execution status is rejected, the action execution result also includes a rejection reason; when the execution status is abnormal execution, the action execution result also includes an abnormal execution reason.

[0083] The post-execution status summary includes the post-execution feeder load status, post-execution phase-to-phase balance status, post-execution cabinet temperature rise status, post-execution alarm status, post-execution cabinet thermal constraint status, and post-execution switch action cooling status.

[0084] The execution status includes one of the following: executed, rejected, abnormal execution, and suspended.

[0085] The sequence version identifier, along with the action execution result and post-execution status summary, are written together into the local execution record.

[0086] When the edge control node determines that the received progressive load balancing sequence does not match the cabinet-level energy efficiency off-load status, the edge control node retains the delayed access action and the prohibited operation action, suspends the recovery action, and filters the disconnection action according to the edge security envelope.

[0087] After receiving the action execution result and post-execution status summary, the cloud control platform writes the abnormal execution reason, rejection reason, and post-execution status summary into the cabinet group load eccentricity map with the corresponding map version identifier, and generates a new progressive load balancing sequence based on the updated cabinet group load eccentricity map. The new progressive load balancing sequence uses the target actions corresponding to the rejection status and abnormal execution status as the basis for the cloud control platform to select target actions when generating subsequent progressive load balancing sequences.

[0088] When performing version comparison, the edge control node uses the sequence version identifier carried by the progressive load balancing sequence, the corresponding map version identifier, and the record version in the local recording area as the comparison objects. When the record version in the local recording area is later than the data source version corresponding to the map version identifier, or when the cabinet-level energy efficiency off-load status in the local recording area is inconsistent with the triggering basis in the progressive load balancing sequence, the edge control node determines that the cabinet-level energy efficiency off-load status does not match the corresponding cabinet group load off-load map status.

[0089] In a preferred embodiment of Example 1, after receiving the cabinet-level energy efficiency off-load status uploaded by the edge control node, the cloud control platform maps the cabinet-level energy efficiency off-load status to the corresponding node and basic graph edge in the cabinet group load off-load graph according to the switch cabinet identifier, feeder affiliation, load unit identifier, power user identifier and phase identifier in the cabinet load record.

[0090] The basic graph edges include power supply attribution edges between switchgear nodes and feeder nodes, access attribution edges between feeder nodes and load unit nodes, power consumption attribution edges between load unit nodes and main power-consuming nodes, phase attribution edges between load unit nodes and phase nodes, as well as edges related to the same switchgear, the same feeder, the same main power-consuming entity, and the same phase. The cloud control platform uses a multi-state marking method to write association results on the basic graph edges, enabling the same basic graph edge to simultaneously retain one or more of the following markers: heavy load association, light load association, temperature rise association, accessible association, disconnectable association, and prohibited operation association. The state source of each marker is recorded as the corresponding edge control node, acquisition time, and recording version.

[0091] The cloud control platform uses the following tag vector to represent the association labels of the basic graph edges:

[0092] ,

[0093] in, Indicates time Next node With nodes The associated marker vectors of the basic graph edges between them; Indicates time Next node With nodes Overloaded association markers for the basic graph edges between them; Indicates time Next node With nodes Lightly loaded association markers for the basic graph edges between them; Indicates time Next node With nodes Temperature rise correlation markers between the basic graph edges; Indicates time Next node With nodes Accessible association markers for the basic graph edges between them; Indicates time Next node With nodes The tangentially separable association markers of the basic graph edges between them; Indicates time Next node With nodes Forbidden association markers for edges in the basic graph; nodes and nodes These are two nodes connected by the edge of the foundation diagram in the load distribution diagram of the cabinet group. This indicates the data collection time corresponding to the cabinet-level energy efficiency off-center load state. A value of 1 for each association flag indicates that the corresponding association is established, and a value of 0 indicates that the corresponding association is not established.

[0094] For the power supply attribution edge between switchgear nodes and feeder nodes, the access attribution edge between feeder nodes and load unit nodes, and the edge related to the same feeder, the cloud control platform writes heavy load associations and light load associations based on the feeder load status:

[0095] ,

[0096] ,

[0097] in, Represents a node With nodes The basic graph edges between them; Indicates time The set of basic graph edges that have a mapping relationship with the feeder load status; Indicates time Below is the basic graph edge The feeder load status of the mapped feeder nodes; Indicates the basis of the graph edges Determine the mapping function for the feeder nodes; This indicates an indicator function; it takes a value of 1 when the condition within the parentheses is true, and a value of 0 when the condition within the parentheses is false. The feeder load status is determined by the edge control nodes according to the feeder load interval. The cloud control platform only writes this status to the basic graph edges belonging to the same feeder, and uses this status as the data source for association markers.

[0098] For the power supply attribution edge between switchgear nodes and feeder nodes, the same-cabinet relationship edge between switchgear nodes and load unit nodes, as well as the same-phase relationship edge and same-power-consumption entity relationship edge related to the switchgear node, the cloud control platform writes the temperature rise correlation based on the cabinet's thermal constraint status:

[0099] ,

[0100] in, Indicates time The set of basic graph edges that have a mapping relationship with the thermal constraint state of the cabinet; Indicates time Below is the basic graph edge The thermal constraint state of the switchgear node obtained by mapping; Indicates the basis of the graph edges Determine the mapping function for the switch cabinet nodes. When the cabinet's thermal constraint status is in the accessible state, the temperature rise correlation flag is set to 0; when the cabinet's thermal constraint status enters the restricted access state, prohibited operation state, or only disconnection state, the temperature rise correlation flag is set to 1, and the cabinet's temperature rise status and alarm status are retained as the status source.

[0101] When the basic graph edge When not directly connected to the switchgear node, The feeder affiliation of the load unit associated with the basic map edge is traced back to the corresponding switchgear node; if the same basic map edge cannot be traced back to a unique switchgear node, then the basic map edge will not be written with accessible association and temperature rise association, and the status source will be marked as mapping anomaly.

[0102] For the access and attribution edges between feeder nodes and load unit nodes, the power consumption attribution edges between load unit nodes and main power consumption nodes, and the phase attribution edges between load unit nodes and phase nodes, the cloud control platform writes accessable associations, disconnectable associations, and prohibited associations based on the load switchable status and cabinet thermal constraint status:

[0103] ,

[0104] ,

[0105] ,

[0106] in, Indicates time The set of basic graph edges that have a mapping relationship with the switchable state of the load; Indicates time Below is the basic graph edge The access permission flag of the mapped load unit; Indicates the basis of the graph edges Determine the mapping function for the load unit nodes; Indicates time Below is the basic graph edge The cut-off permission flag for the mapped load unit; Indicates time Below is the basic graph edge The switching lockout flag of the mapped load unit. The access permission flag is formed by the load attribute that can be delayed or the load attribute that can be restored; the disconnection permission flag is formed by the load attribute that can be interrupted; the switching lockout flag is formed by the load attribute that can be guaranteed or the load attribute that restricts switching.

[0107] When the cabinet's thermal constraint state is in a prohibited state, the basic map edge related to the corresponding switch cabinet node is written with the prohibited association; when the load unit has a switching lockout flag, the basic map edge related to the load unit node is written with the prohibited association.

[0108] After completing edge marking, the cloud control platform aggregates the associated markings for the same switchgear node and its associated feeder nodes. For switchgear nodes... When a base graph edge associated with a switchgear node or its feeder node acquires at least one of a heavy load association or a temperature rise association, the switchgear node is marked as a release object; when both a heavy load association and a temperature rise association are established, the switchgear node is designated as a priority release object within the same action level.

[0109] ,

[0110] in, Indicates time Lower switchgear node The marker to be released; This indicates the switchgear nodes in the load distribution diagram of the switchgear group; Represents a node With nodes The basic graph edges between them; Indicates time Downstream and switchgear node and the set of basic graph edges related to its feeder nodes; The definition of overloaded association tags will be used as described above; Using the definition of the aforementioned temperature rise correlation marker; node and nodes This refers to two nodes connected by an edge in the base diagram within the load distribution map of the switchgear group. When the "to be released" flag is set to 1, the cloud control platform includes this switchgear node as a target for release when generating the progressive load balancing sequence; when the "to be released" flag is set to 0, this switchgear node is not included in the target set. When both values ​​are 1, the cloud control platform simultaneously writes both the overload association and the temperature rise association into the trigger criteria for this switchgear node.

[0111] When both accessible and prohibited associations are satisfied on the same basic graph edge, the cloud control platform uses the prohibited association as a constraint for action generation, while retaining the accessible association as a source of historical status. When both disconnectable association and switching lock flags are present on the same load unit node, the cloud control platform uses the prohibited association to constrain the disconnection action, excluding the load unit from the candidate range for releasing interruptible loads. Through the above processing, the association markers on the basic graph edge can simultaneously reflect feeder load off-center loading, cabinet temperature rise constraints, and load switching availability, providing a graph-based basis for subsequently determining the objects to be released, the objects that can be carried, and generating a progressive load balancing sequence.

[0112] Example 2:

[0113] Based on Example 1, this example provides a specific method for generating delayed access actions, prohibited access actions, disconnection actions, and recovery actions in the cloud-edge collaborative control method for optimizing power distribution load energy efficiency;

[0114] When the cloud control platform generates a progressive load balancing sequence, it forms a set of objects to be released and a set of objects that can be carried based on the load eccentricity map of the cabinet group used when the sequence is generated, and reads the recovery candidate set that matches the map version identifier of the load eccentricity map of the cabinet group.

[0115] The set of objects to be released includes switchgear identifier, feeder affiliation, association type, status source, and load unit range. The association type includes heavy load association and temperature rise association. The set of objects that can be carried includes switchgear identifier, feeder affiliation, light load association, prohibited operation association mark, and operation permission status. The set of candidates for recovery includes load unit identifier, target switchgear, target feeder, recovery source, most recent disconnection action, and cabinet thermal constraint status.

[0116] Furthermore, a new access request refers to a load unit access application received by the cloud control platform or a load unit access event reported by the edge control node. Its fields include request identifier, target switchgear, target feeder, load unit, submission time, and load attributes. Recovery source refers to the data source corresponding to the disconnection action, abnormal execution status, or pause status when a load unit enters the recovery candidate set. Status sources include record version, map version identifier, acquisition time, and the cabinet load record on which the corresponding status is based.

[0117] In one implementation, the cloud control platform generates a progressive load balancing sequence in the order of restricting new access, releasing interruptible loads, and restoring recoverable loads; the cloud control platform first processes new access requests based on the set of objects to be released, and generates delayed access actions or prohibited access actions;

[0118] When the heavy load association or temperature rise association corresponding to the object to be released still exists, the cloud control platform generates a disconnection action based on the load switchable status.

[0119] When the cloud control platform determines that the overload association and temperature rise association corresponding to the object to be released have been released based on the action execution result and the post-execution status summary, the cloud control platform updates the release mark of the object to be released to the released status and generates a recovery action based on the recovery candidate set.

[0120] Each action in the progressive load balancing sequence includes the target switchgear, target feeder, load unit, action type, action sequence, and triggering basis.

[0121] Delayed access action refers to the cloud control platform reserving a new access request as a subsequent access candidate and temporarily suspending the generation of access control. Disallow access action refers to the cloud control platform marking the target switchgear, target feeder, or load unit as prohibited from new access. Disconnection action refers to the control action by which the edge control node disconnects the target load unit from the power supply connection. Restoration action refers to the control action by which the edge control node reconnects a disconnected target load unit with restoreable load attributes to the power supply connection.

[0122] In a preferred embodiment of Example 2, when the cloud control platform generates a progressive load balancing sequence, it uses the objects to be released and the objects that can be carried in the load balancing map of the cabinet group as constraint boundaries, and generates actions according to the action hierarchy of restricting new access, releasing interruptible loads, and restoring recoverable loads.

[0123] When generating each round of sequences, the cloud control platform reads the heavy load association, light load association, temperature rise association, cabinet thermal constraint status, load switchable status, and switch action cooling status in the cabinet group load off-center load map corresponding to the current map version identifier. It also prioritizes new access requests that have not yet been executed in the current round, making new access requests that target switch cabinets that are pending release, or target switch cabinets that are in a restricted access state, prohibited operation state, or only allowed disconnection state, and converts them into delayed access actions or prohibited operation actions.

[0124] The heavy load correlation, light load correlation, temperature rise correlation, cabinet thermal constraint status, load switchability status, and switch action cooling status used in the current round are all taken from the cabinet group load off-center load map corresponding to the same map version identifier; data collection time Used to identify the source of the state and the time of sequence generation. Used to identify the generation time of the action sequence. If the record versions corresponding to the same target switchgear, target feeder, or load unit are inconsistent, the cloud control platform retains the delayed access action and the prohibited operation action, suspends the recovery action, and only generates the disconnection action when the candidate permission flag of the disconnection action can be verified by the same record version.

[0125] For new access requests, the cloud control platform generates restrictions on new access actions based on the target switchgear's pending release flag, cabinet thermal constraint status, load attributes, and switch action cooling status, and matches them from top to bottom according to the following conditions; if a previous condition is met, the next condition will not be matched:

[0126] like ,but ,

[0127] like ,but ,

[0128] like ,but ,

[0129] in, Indicates the sequence generation time Add new access request The corresponding access processing result; Indicates the sequence generation time Add new access request The thermal constraint state of the corresponding target switchgear cabinet; Indicates a new access request The corresponding target switchgear; Indicates the sequence generation time Add new access request The switching lockout flag for the corresponding load unit; Indicates the sequence generation time Add new access request The target switchgear is marked as to be released; Indicates the sequence generation time Add new access request The corresponding target switchgear's switching action cooling indicator. A switching lockout indicator value of 1 indicates that the corresponding load unit has a load supply guarantee attribute or a load switching restriction attribute; a value of 0 indicates that the corresponding load unit is not in a switching lockout state.

[0130] Used in the sequence generation stage and This is the switching lock flag under the load unit index caliber, and its status originates from the map marking stage. The same; the difference is, The load cells obtained by edge mapping of the basic graph are used as indices. Using the load unit corresponding to the new access request as the index, Use the load unit for the action to be generated as the index.

[0131] When the switch action cooling flag is set to 1, it indicates that the target switch cabinet is in a cooling hold state; when it is set to 0, it indicates that the target switch cabinet is in an action permit state.

[0132] After processing the new access request restrictions, the cloud control platform reviews the objects to be released. If the overload or temperature rise associations corresponding to the objects to be released still exist, load units with interruptible load attributes, no guaranteed load attributes, no restricted load switching attributes, and whose switching action cooling state is in an action-permitted state are selected from the load units to which the objects to be released belong, as candidates for disconnection actions. The permission criteria for disconnection action candidates are:

[0133] ,

[0134] in, Indicates the sequence generation time Lower load unit Candidate license flag for the cut-off action; Indicates the sequence generation time Lower load unit The target switchgear's pending release marker; Represents load unit The target switchgear; Indicates the sequence generation time Lower load unit Interruptible load attribute flag; Indicates the sequence generation time Lower load unit The throwing and locking mark; Indicates the sequence generation time Lower load unit Cooling indicator for the switchgear of the target switchgear; This indicates an indicator function; the value is 1 when the condition within the parentheses is true, and 0 when the condition within the parentheses is false. The interruptible load attribute flag is 1 when the load unit can be released, and 0 when the load unit is not. The disconnect action candidate permission flag is 1 when the cloud control platform can generate a disconnect action based on the load unit, and 0 when no disconnect action is generated.

[0135] For multiple shelving action candidates, the cloud control platform prioritizes them according to their contribution to load release, temperature rise mitigation, and inter-phase balance improvement, allowing load units that significantly alleviate heavy load and temperature rise correlations to enter the gradual load balancing sequence first.

[0136] ,

[0137] in, Indicates the sequence generation time Lower load unit The priority value for cutting off; Indicates the sequence generation time The weight of the contribution of load release; Indicates the sequence generation time Cut-off load unit Expected feeder load release; Indicates the sequence generation time Lower load unit The maximum load limit of the target feeder; Represents load unit The target feeder belongs to; Indicates the sequence generation time The weight of the contribution of the temperature rise mitigation; Indicates the sequence generation time Cut-off load unit The expected reduction in cabinet temperature rise; Indicates the sequence generation time Lower load unit Temperature rise limit of the target switchgear; Indicates the sequence generation time The weight of the contribution of the lower phase equilibrium improvement; Indicates the sequence generation time Cut-off load unit The expected improvement in phase balance is as follows: The load limit is jointly determined by the feeder's rated capacity and feeder operating margin; the temperature rise limit is determined by the temperature rise threshold corresponding to the cabinet's thermal constraint state; the weights of the load release contribution, the temperature rise mitigation contribution, and the phase balance improvement contribution are all greater than or equal to 0 and less than or equal to 1, and the sum of the three is 1.

[0138] The expected release of feeder load is based on load unit. The average load value is calibrated within a preset number of valid collection periods prior to the current recording version's corresponding collection time. A valid collection period is defined as a collection period where communication is normal and load records are complete. If the number of valid collection periods is lower than a preset lower limit... The value is set to 0. The expected reduction in cabinet temperature rise is determined based on the correspondence between the load decrease and the temperature rise decrease in the target switchgear's historical operation records. This is applied when historical operation records are insufficient, operation results are abnormal, or cabinet temperature rise status is missing. The value is 0. The phase balance improvement is determined according to the percentage decrease in the maximum load difference between phases before and after the phase disconnection, and the value is greater than or equal to 0 and less than or equal to 1; if the maximum load difference between phases does not decrease after the phase disconnection, or if the phase identifier is missing, The value is 0. Both the load limit and temperature rise limit are positive.

[0139] When calculating the priority value for cutting off, and All are truncated to a range greater than or equal to 0 and less than or equal to 1; when or When a value is missing, abnormal, or cannot be confirmed by the current record version, the corresponding contribution item is set to 0. The weights of load release contribution, temperature rise mitigation contribution, and phase balance improvement contribution are calibrated offline by the cloud control platform according to the operating scenario and bound to the sequence version identifier of the current progressive load balancing sequence; when the weights are missing or the sum of the three is not 1, , and All are conservatively replaced with 1 / 3. The larger the priority value of the cut-off action, the earlier the cut-off action is arranged in the same action level.

[0140] When the action execution result and post-execution status summary uploaded by the edge control node show that the overload association and temperature rise association corresponding to the object to be released have been released, the cloud control platform updates the release marker to the released state and generates a recovery action from the recovery candidate set. Recovery actions are only generated for load units with recoverable load attributes and are constrained by the load-bearing object, cabinet thermal constraint status, and switch action cooling status. The load units in the recovery candidate set are all written by the edge control node according to the recoverable load attributes after the action is executed; load units without recoverable load attributes do not enter the recovery candidate set.

[0141] ,

[0142] in, Indicates the sequence generation time Lower load unit Recovery action candidate license flag; Indicates the sequence generation time The following is a set of recovery candidates uploaded by the edge control node and confirmed by the cloud control platform; Indicates the sequence generation time Lower load unit The target switchgear's pending release marker; Indicates the sequence generation time Lower load unit The thermal constraint status of the target switchgear cabinet; Indicates the sequence generation time Lower load unit Cooling indicator for the switchgear of the target switchgear; Indicates the sequence generation time Lower load unit The load-bearing flag for the target feeder. When the recovery action candidate permission flag is set to 1, the cloud control platform can generate a recovery action based on the load unit; when the value is 0, the load unit remains in the recovery candidate set or is removed from the recovery candidate set according to the post-execution state summary of the edge control node. A load-bearing flag value of 1 indicates that the target feeder has a light load association and is not marked with a prohibited association; a value of 0 indicates that the target feeder is not a recovery load-bearing object.

[0143] When the value of the release flag is 0, the corresponding object to be released is in a released state; when the value of the release flag is 1, the corresponding object to be released is in a pending release state.

[0144] When using the load capacity flag to determine the permission for recovery actions, the calibration calibrates based on the assumption that the target feeder has not yet returned to a heavy-load state after the estimated recovery load of the current load unit $s$ is included. The estimated recovery load is calibrated according to the load record corresponding to the load unit before it enters the recovery candidate set. If there are multiple valid load records, the average load within the valid acquisition period before the acquisition time corresponding to the current record version is used. If the corresponding load record is missing, the communication status is abnormal, or it cannot correspond to the current record version, the estimated recovery load is treated as missing. When there are multiple recovery candidates for the same target feeder, the cloud control platform forms a pre-arranged order according to the sorting value of the recovery actions within the same action level from high to low, and accumulates the estimated recovery load item by item according to this pre-arranged order. After accumulation, the load capacity flag is re-determined. If the estimated recovery load is missing, the feeder load status is missing after the execution of the target feeder, or the target feeder returns to a heavy-load state after accumulation, the load capacity flag for the corresponding recovery action is treated as 0.

[0145] The cloud control platform hierarchically arranges the generated delayed access actions, prohibited access actions, disconnection actions, and recovery actions, placing the action to restrict new access before the action to disconnection, and the action to disconnection before the action to recovery. Within the same level, disconnection actions are arranged in descending order of disconnection priority, and recovery actions are arranged in descending order of the retention time of the corresponding target switchgear.

[0146] ,

[0147] in, Represents the first in the progressive load balancing sequence One candidate action; Represents the first in the progressive load balancing sequence One candidate action; This indicates the order of precedence among candidate actions; Indicates the sequence generation time Next candidate action Action level value; Indicates the sequence generation time Next candidate action Action level value; Indicates the sequence generation time Next candidate action The sorting value within the same action level; Indicates the sequence generation time Next candidate action The ranking value within the same action level. The action level value for delayed access and disabling actions is set to 1, the action level value for disconnection actions is set to 2, and the action level value for recovery actions is set to 3. When the candidate action is a disconnection action, its ranking value adopts the disconnection priority value; when the candidate action is a recovery action, its ranking value adopts the holding time of the target switchgear since the most recent switching action (described later); when the candidate action is a delayed access action or a disabling action, its ranking value adopts the priority of the submission time of the new access request. The priority of the submission time of the new access request is set in descending order of submission time, with earlier submission times resulting in higher ranking values.

[0148] The retention time of the target switchgear after the most recent switching action is based on the execution time in the action execution result of the local execution record and the corresponding sequence version identifier in the cloud. When the record of the most recent switching action is missing, the time when the target switchgear enters the action permission state in the current record version is used as the starting point of the retention time. If the time of entering the action permission state cannot be confirmed, or the execution result cannot be confirmed, the communication status is abnormal, or the switching action cooling status is missing, the generation of the recovery action of the target switchgear is suspended.

[0149] Each action in the progressive load balancing sequence is recorded in the target switchgear, target feeder, load unit, action type, action sequence, and triggering basis. Among the triggering basis, delayed access actions record at least one of the following: pending release flag, restricted access status, or cooling hold status; prohibited operation actions record at least one of the following: prohibited operation status, guaranteed load attribute, or restricted load switching attribute; disconnection actions record at least one of the following: overload association or temperature rise association, interruptible load attribute, and action permission status; recovery actions record at least one of the following: recovery candidate set source, bearable flag, accessible status, and action permission status.

[0150] The above action generation process is triggered by each update of the cabinet group load off-center map or by the upload of action execution results by the edge control node. If the cloud control platform does not receive the action execution result with the corresponding sequence version identifier, or if the post-execution status summary lacks the post-execution feeder load status, post-execution cabinet temperature rise status, or post-execution switch action cooling status, the cloud control platform will not generate a recovery action and will limit the disconnection action to the load units that have met the disconnection action candidate permission flag.

[0151] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit them. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some or all of the technical features therein. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of this application.

[0152] Furthermore, those skilled in the art will understand that although some embodiments herein include certain features included in other embodiments but not others, combinations of features from different embodiments are meant to be within the scope of this application and form different embodiments. For example, all the embodiments above can be used in any combination. The information disclosed in this background section is intended only to enhance the understanding of the general background of this application and should not be construed as an admission or in any way implying that such information constitutes prior art known to those skilled in the art.

Claims

1. A cloud-edge collaborative control method for optimizing power distribution load energy efficiency, applied to a power distribution system including a cloud control platform, edge control nodes, and intelligent switchgear, characterized in that, include: The edge control node collects cabinet-end operation data and generates cabinet load records and cabinet-level energy efficiency off-load status. The cloud control platform constructs a load distribution map of the cabinet group based on the cabinet load records and the cabinet-level energy efficiency off-load status. The load distribution map of the cabinet group includes switch cabinet nodes, feeder nodes, load unit nodes and basic map edges. The cloud control platform writes the feeder load status, cabinet thermal constraint status, and load switchability status into the basic graph edge, forming at least one association marker among heavy load association, light load association, temperature rise association, and no-operation association containing the status source; When there are objects to be released marked by at least one of heavy load association and temperature rise association in the load distribution diagram of the cabinet group, and there are load-bearing objects marked by light load association but not marked by prohibited loading association, the cloud control platform generates a progressive load balancing sequence based on the objects to be released, load-bearing objects and the thermal constraint status of the cabinet. The edge control node forms an edge safety envelope based on the cabinet temperature rise status, alarm status, communication status, switch action cooling status, and load switchability status. It then filters the actions in the progressive load balancing sequence, executes the filtered actions, and uploads the action execution results and post-execution status summary. The cloud-based control platform updates the load eccentricity map of the cabinet group based on the action execution results and the post-execution status summary, and uses the updated load eccentricity map of the cabinet group as the basis for generating the subsequent progressive load balancing sequence.

2. The cloud-edge collaborative control method for optimizing power distribution load energy efficiency according to claim 1, characterized in that, Cabinet operation data includes current, voltage, power, switch status, cabinet temperature rise status, alarm status, and communication status; cabinet load records include switch cabinet identification, feeder affiliation, load unit identification, power user identification, phase identification, data collection time, and load attributes. A load unit is an electrical load or a collection of electrical loads that is connected to the intelligent switch cabinet feeder and can be identified as an object to be connected, disconnected, or restored. Edge control nodes collect cabinet operation data within the same acquisition cycle according to switch cabinet identification, feeder affiliation, and load unit identification.

3. The cloud-edge collaborative control method for optimizing power distribution load energy efficiency according to claim 2, characterized in that, Load attributes include at least one of the following: guaranteed load attributes, interruptible load attributes, delayed access load attributes, recoverable load attributes, and restricted switching load attributes; The edge control node writes the load attributes into the corresponding load unit identifier and saves the cabinet load record in the local recording area; the local recording area saves the record version of the cabinet load record according to the switch cabinet identifier and the time of acquisition; When communication between the edge control node and the cloud control platform fails, the edge control node verifies the received progressive load balancing sequence using the cabinet load record in its local recording area.

4. The cloud-edge collaborative control method for optimizing power distribution load energy efficiency according to claim 1, characterized in that, When the edge control node generates the cabinet-level energy efficiency off-load state, it determines the feeder load state according to the feeder load range, the phase balance state according to the load difference of each phase, the cabinet thermal constraint state according to the cabinet temperature rise state and alarm state, the switch action cooling state according to the holding time after the most recent switch action, and the load switchable state according to the load attributes. The cabinet's thermal constraint status includes accessible status, restricted access status, prohibited operation status, and disconnection-only status; the switch operation cooling status includes cooling hold status and operation permission status.

5. The cloud-edge collaborative control method for optimizing power distribution load energy efficiency according to claim 1, characterized in that, The load distribution diagram of the cabinet group also includes the main power consumption node and the phase node; The basic diagram edges include the power supply attribution edge between switch cabinet nodes and feeder nodes, the access attribution edge between feeder nodes and load unit nodes, the power consumption attribution edge between load unit nodes and power-consuming entity nodes, the phase attribution edge between load unit nodes and phase nodes, the same cabinet relationship edge, the same feeder relationship edge, the same power-consuming entity relationship edge, and the same phase relationship edge. The cloud control platform writes the feeder load status into the basic graph edge related to the feeder node, writes the cabinet thermal constraint status into the basic graph edge related to the switch cabinet node, and writes the load switchability status into the basic graph edge related to the load unit node.

6. The cloud-edge collaborative control method for optimizing power distribution load energy efficiency according to claim 1, characterized in that, When the cloud control platform generates a progressive load balancing sequence, it forms a set of objects to be released and a set of objects that can be carried based on the load eccentricity map of the cabinet group used when the sequence is generated, and reads the recovery candidate set that matches the map version identifier of the load eccentricity map of the cabinet group. The set of objects to be released includes switch cabinet identifier, feeder affiliation, association type, status source, and load unit range. The association type includes overload association and temperature rise association. The set of load-bearing objects includes switchgear identification, feeder affiliation, light load association, prohibited operation association mark, and action permission status. The set of recovery candidates includes load unit identification, target switchgear, target feeder, recovery source, most recent disconnection action, and cabinet thermal constraint status.

7. The cloud-edge collaborative control method for optimizing power distribution load energy efficiency according to claim 6, characterized in that, The cloud control platform generates a progressive load balancing sequence in the order of restricting new access, releasing interruptible loads, and restoring recoverable loads; The cloud control platform first processes new access requests based on the set of objects to be released, and generates delayed access actions or disallow access actions; When the heavy load association or temperature rise association corresponding to the object to be released still exists, the cloud control platform generates a disconnection action based on the load switchable status. When the cloud control platform determines that the overload association and temperature rise association corresponding to the object to be released have been released based on the action execution result and the post-execution status summary, the cloud control platform updates the object to be released to the released status and generates a recovery action based on the recovery candidate set; Each action in the progressive load balancing sequence includes the target switchgear, target feeder, load unit, action type, action sequence, and triggering basis.

8. The cloud-edge collaborative control method for optimizing power distribution load energy efficiency according to claim 1, characterized in that, When the edge control node performs screening, it will reject new access actions if the target switchgear has abnormal temperature rise or the alarm has not been cleared; it will reject actions again if the target switchgear or target feeder is in the cooling state of switching operation; it will reject disconnection actions if the target load unit has load supply protection attributes or load switching restriction attributes; it will reject target actions if the target action changes the phase balance state to an unfavorable state; and it will reject access actions, disconnection actions, and recovery actions if the target switchgear communication state is abnormal and the action result cannot be confirmed. The edge control node generates rejection reasons for the rejected target actions.

9. The cloud-edge collaborative control method for optimizing power distribution load energy efficiency according to claim 1, characterized in that, When the edge control node executes the filtered action, it performs the access, disconnection or restoration control item by item in the order of the progressive load balancing sequence. After each action is executed, it reads the switch status, cabinet temperature rise status, alarm status, feeder load status and phase balance status of the target switch cabinet and generates a post-execution status summary based on the reading results. When the execution result of the action is inconsistent with the target action, the edge control node stops executing subsequent actions of the same target switch cabinet or the same target feeder and generates an abnormal execution reason; when the thermal constraint status of the cabinet after execution is an accessible state and the target load unit has a recoverable load attribute, the edge control node adds the target load unit to the recovery candidate set.

10. The cloud-edge collaborative control method for power distribution load energy efficiency optimization according to claim 9, characterized in that, The cloud-based control platform sets a version identifier for the load distribution diagram of the switchgear group; the action execution results uploaded by the edge control nodes include the sequence version identifier, target switchgear, target feeder, load unit, action type, execution status, and execution time. The execution status includes one of the following: executed, rejected, abnormal execution, and suspended. The rejected status corresponds to the reason for rejection, and the abnormal execution status corresponds to the reason for abnormal execution. The post-execution status summary includes the post-execution feeder load status, post-execution phase-to-phase balance status, post-execution cabinet temperature rise status, post-execution alarm status, post-execution cabinet thermal constraint status, and post-execution switch action cooling status. The cloud control platform writes the action execution results and post-execution status summary into the cabinet group load eccentricity map with the corresponding map version identifier, and uses the target actions corresponding to the rejected state and abnormal execution state as the basis for the cloud control platform to select target actions when generating subsequent progressive load balancing sequences.

Images