Event-triggered negotiation-based power control method for charging device cluster and charging system

Abstract

The application discloses an event-triggered negotiation type power control method of a charging device cluster and a charging system, relates to the field of electric vehicle charging, and comprises the following steps: detecting the heartbeat state of a total control node, and performing mode switching between a centralized coordination mode and a distributed autonomous mode according to the heartbeat state; performing trigger event detection in the running mode after switching; maintaining each device in a zero negotiation silent state when no trigger event is detected; starting distributed negotiation when at least one type of trigger event is detected, iteratively updating each device by taking the reducible proportion of the current power relative to the demand power as a consistency variable, and stopping until the consistency variable of each device converges to a global consistent value; determining the upper limit of the constrained power of each device according to the global consistent value after convergence; and performing power control according to the upper limit of the constrained power. The application can avoid the paralysis of charging pile cluster scheduling caused by the failure of the total controller, and reduce the periodic invalid communication overhead.

Description

Technical Field

[0001] This application relates to the field of electric vehicle charging, and in particular to an event-triggered negotiation power control method and charging system for a cluster of charging devices. Background Technology

[0002] Currently, in charging stations where multiple charging piles share transformer capacity, a centralized control architecture is conventionally used for load management and power scheduling. The central controller collects the operating data of all charging piles in the station, calculates the allowable power of each pile based on the upper limit of the transformer capacity, and issues scheduling instructions to each charging pile through a power allocation strategy. Each charging pile acts as a passive actuator and outputs power according to the instructions.

[0003] In related technologies, the operation of each charging pile in a centralized control architecture relies entirely on the central controller. When the central controller experiences hardware failure, software anomalies, or communication interruptions, all charging piles at the station immediately lose their dynamic scheduling capabilities. Existing solutions typically revert to the preset fixed offline power operation of each pile after the central controller fails, resulting in the transformer capacity not being dynamically utilized, some online piles being forced to reduce power or shut down, and a significant decrease in charging efficiency. Furthermore, this architecture relies on the central controller to periodically collect data from the entire station and issue commands. Regardless of changes in the charging station's load status, periodic communication continues uninterrupted, generating a large amount of invalid communication overhead during steady-state operation, leading to excessive load on the communication bus and severely affecting the real-time performance of other control messages.

[0004] Therefore, how to avoid the paralysis of charging pile cluster scheduling due to the failure of the main controller, and at the same time reduce the periodic invalid communication overhead, is an urgent technical problem to be solved. Summary of the Invention

[0005] The main purpose of this application is to provide an event-triggered negotiation power control method and charging system for a charging equipment cluster, so as to avoid the paralysis of the charging pile cluster scheduling due to the failure of the main controller, and at the same time reduce the periodic invalid communication overhead.

[0006] To achieve the above objectives, this application provides an event-triggered negotiation-based power control method for a charging device cluster, applied to a charging system. The charging system includes at least two power modules and at least one charging interface, with the power modules connected via a controllable switch. The method includes: Detect the heartbeat status of the central control node, and switch the operating mode between centralized coordination mode and distributed autonomous mode based on the heartbeat status; Trigger event detection is performed in the switched operating mode; When no triggering event is detected, all devices remain in a zero-negotiation silent state; When at least one type of triggering event is detected, distributed negotiation is initiated, and each device is iteratively updated using the current power as a percentage of the power required to be reduced as a consistency variable until the consistency variable of each device converges to a globally consistent value. Based on the converged global consensus value, determine the constrained power limit for each device; Power control is performed based on the power upper limit after the constraints are applied.

[0007] Optionally, the step of switching the operating mode between centralized coordination mode and distributed autonomous mode based on the heartbeat state includes: When the number of consecutive periods without receiving a heartbeat message from the central control node reaches a first preset threshold, the system switches from the centralized coordination mode to the distributed autonomous mode. When the number of consecutive heartbeat messages received from the central control node in the distributed autonomous mode reaches a second preset threshold, the system switches back from the distributed autonomous mode to the centralized coordination mode. The mode switching is performed independently by each device, and the power output is not interrupted during the switching process.

[0008] Optionally, the triggering events include the following four categories: The total power limit over event is triggered when the real-time total power of all devices exceeds the preset detection threshold of the total capacity limit, or when the real-time total power recovers to the preset exit threshold of the total capacity limit; The single device power deviation event is triggered when the deviation between the current power of any device and the negotiated upper limit of power exceeds a preset deviation threshold. Topology state change events are triggered when a new device is connected, a device is disconnected, or the change in the power demand of a device exceeds a preset threshold. The communication loss event is triggered when the heartbeat messages of a neighboring device time out a preset number of times consecutively.

[0009] Optionally, the method further includes: The preset detection threshold for the total power over-limit event and the preset deviation threshold for the single device power deviation event are dynamically adjusted based on the total load rate. When the total load rate is higher than the preset high load threshold, the preset detection threshold and the preset deviation threshold are reduced. When the total load rate is lower than the preset low load threshold, the preset detection threshold and the preset deviation threshold are increased.

[0010] Optionally, the step of initiating distributed negotiation, using the proportion of current power reduction relative to demand power as a consistency variable to iteratively update each device until the consistency variable of each device converges to a globally consistent value, includes: Each device sends its consistency variables to neighboring devices and receives the consistency variables from the neighboring devices. Assign weight values ​​to all neighboring devices of each device, wherein the weight values ​​are determined based on the neighbor relationships of the devices in the physical topology, and the weight between devices that are not neighbors is zero; Based on the weight of each device and the weight of each neighboring device, the consistency variables of each device and the received consistency variables of the neighboring devices are weighted and merged, and the consistency variables of each device are updated. Repeat the above sending, receiving, and weighted merging process until the deviation between the consistency variables of all devices is less than the preset convergence accuracy.

[0011] Optionally, the weight value is determined as follows: The weight assigned to the neighboring device by the current device is inversely proportional to the maximum of the number of neighbors of the current device and the number of neighbors of the neighboring device; The weight assigned to the current device is 1 minus the sum of the weights assigned to all the neighboring devices.

[0012] Optionally, determining the constrained power upper limit for each device based on the converged global consensus value includes: The three-level constraint power boundary is calculated in a nested manner according to priority. The first priority is the upper limit of the required power of each device, the second priority is the upper limit of the safe power of each device, and the third priority is the upper limit of the total capacity power of all devices. The product of the globally consistent value and the power requirement limit of each device is used as the constrained power limit; The upper limit of the constrained power is verified by the three-layer constrained power boundary. When the total capacity power limit in the three-layer constraint is triggered, calculate the average load reduction amount that each device can bear under the demand power limit constraint and the safe power limit constraint. If the load reduction margin of all equipment is sufficient to bear the average load reduction amount, then the average load reduction shall be performed on each piece of equipment. If the load reduction margin of some equipment is insufficient to bear the equally distributed load reduction amount, the load reduction amount shall be allocated in a weighted manner according to the load reduction margin of each equipment. If the total power capacity is still exceeded after weighted allocation, overload protection will be triggered, reducing the power of each device to the preset protection ratio.

[0013] Optionally, the method further includes: Detect the online status of neighboring devices and mark devices that have not received heartbeat messages for a continuous period of time as offline devices; Monitor the instantaneous power of the offline device during actual operation; Obtain the estimated power of the offline device, and calculate the release capacity based on the difference between the estimated power and the instantaneous power; The releasable capacity is added to the allocable capacity pool and allocated to each online device according to the power demand ratio of each online device.

[0014] Optionally, the method further includes: Is the detection equipment cluster not in a fully configured state? When the device is not fully equipped, the idle capacity value is determined based on the relationship between the total capacity power limit, the sum of the current actual power of all online devices, and the sum of the current actual power of all offline devices. When the idle capacity value is greater than zero, the idle capacity is allocated to each of the online devices according to the proportion of the power demand of each of the online devices to the total power demand of all the online devices.

[0015] In addition, to achieve the above objectives, this application also provides a charging system, including: multiple charging piles, each charging pile being equipped with a local controller and a neighbor communication module, wherein the local controller is used to execute an event-triggered negotiation power control method for a charging equipment cluster as described above, and the neighbor communication module is used to realize data interaction between the charging piles.

[0016] This application proposes an event-triggered negotiation-based power control method and charging system for a charging device cluster. By detecting the heartbeat status of the central control node and switching between centralized coordination mode and distributed autonomous mode based on the heartbeat status, it enables each device to automatically switch to distributed autonomous mode to continue operation when the central control node fails, avoiding dependence on a single central control node. By detecting trigger events in the switched operating mode, each device remains in a zero-negotiation silent state when no trigger event is detected, and distributed negotiation is only initiated when a trigger event is detected, avoiding periodic invalid communication. By iteratively updating each device with the ratio of the current power to the required power reduction as a consistency variable until convergence, it achieves a fair allocation of the load reduction obligation of each device, unaffected by an increase in the number of devices. By determining the constrained power upper limit of each device based on the converged global consensus value and performing power control accordingly, it ensures that the power allocation result meets the physical limits of each device. Because this application allows each device to automatically switch to a distributed autonomous mode to continue distributed negotiation when the central control node fails, without manual intervention or interruption of power output, it fundamentally eliminates the risk of centralized single-point failure. Furthermore, this application only initiates negotiation when a trigger event occurs; otherwise, devices only maintain heartbeat communication. Compared to the periodic polling schemes of existing technologies, this significantly reduces communication overhead and effectively releases communication bus resources. In addition, the iterative update process based on consistency variables eliminates the need for global information, ensuring the scalability of the solution in large-scale device clusters. Therefore, this application achieves the technical effect of reducing periodic invalid communication overhead while avoiding scheduling paralysis caused by a single point of failure in the central control node, solving the technical problems of low reliability and high communication overhead in charging equipment clusters that rely on centralized central control. Attached Figure Description

[0017] Figure 1 This is a scenario example of an event-triggered negotiation power control method for a charging device cluster according to an embodiment of this application.

[0018] Figure 2 This is one of the flowcharts for an event-triggered negotiation power control method for a charging device cluster according to an embodiment of this application.

[0019] Figure 3 This is a second flowchart of an event-triggered negotiation power control method for a charging device cluster according to an embodiment of this application.

[0020] Figure 4 This is the third flowchart of an event-triggered negotiation power control method for a charging device cluster according to an embodiment of this application.

[0021] Figure 5 This is the fourth flowchart of an event-triggered negotiation power control method for a charging device cluster according to an embodiment of this application.

[0022] Figure 6 This is the fifth flowchart of an event-triggered negotiation power control method for a charging device cluster according to an embodiment of this application.

[0023] Figure 7 This is the sixth flowchart of an event-triggered negotiation power control method for a charging device cluster according to an embodiment of this application.

[0024] Figure 8 This is the seventh flowchart of an event-triggered negotiation power control method for a charging device cluster according to an embodiment of this application.

[0025] Figure 9 A schematic diagram of the physical structure of a main controller is provided.

[0026] In the diagram: 110, power module; 120, charging interface; 130, main controller; 140, power distribution device; 150, charging terminal; 910, processor; 920, communication interface; 930, memory; 940, communication bus.

[0027] The realization of the purpose, functional features and advantages of this application will be further explained in conjunction with the embodiments and with reference to the accompanying drawings. Detailed Implementation

[0028] To make the objectives, technical solutions, and advantages of this application clearer, the technical solutions of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0029] In charging stations with multiple charging piles sharing transformer capacity, a centralized control architecture is conventionally used for load management and power scheduling. The central controller acts as the sole scheduling decision node for the entire station, uniformly collecting data from the entire station and issuing scheduling commands. This architecture places extremely high demands on the stability of the central control node and the reliability of the communication link. However, in actual operation, the central controller faces the risk of hardware failure, software anomalies, or communication interruptions. Once the central controller fails, all charging piles in the station will lose their dynamic scheduling capabilities and revert to preset fixed power operation, resulting in the transformer capacity not being dynamically utilized, and some online piles being forced to reduce power or shut down. Simultaneously, this architecture relies on the central controller periodically collecting data from the entire station. Regardless of changes in the station's load status, communication remains continuous, causing significant unnecessary communication overhead during steady-state operation. For large stations, the high communication bus load caused by periodic polling severely affects the real-time performance of other control messages. Furthermore, existing solutions lack a multi-layered constraint nesting system from individual device physical limits to the total system capacity, and lack priority logic for constraint resolution, potentially leading to local violations of safety constraints under complex operating conditions.

[0030] To address this, this application provides an event-triggered negotiation-based power control method and charging system for a charging equipment cluster. By detecting the heartbeat status of the central control node and switching between centralized coordination mode and distributed autonomous mode, the system remains silent when there are no events and triggers negotiation when there are events. It uses the power reduction ratio as a consistency variable for iterative convergence and ensures that power allocation meets the physical limits of individual devices and the total system capacity constraints through a three-layer constraint nesting mechanism. This effectively reduces the periodic invalid communication overhead while avoiding scheduling paralysis caused by single-point failures. It solves the technical problems of low reliability, high communication overhead, and lack of multi-level security constraint protection in charging equipment clusters that rely on centralized central control.

[0031] For ease of understanding, this specification provides a scenario example of an event-triggered negotiation-based power control method for a charging device cluster. This scenario example is applied to applications such as... Figure 1 The example application environment is shown. Specifically, this scenario is an application scenario for a large electric vehicle charging station. In this scenario example, the charging system of the charging station may include at least two power modules 210, at least one charging interface 220, a main controller 230, and a power distribution device 240.

[0032] It should be noted that the charging system in this scenario example is a distributed charging system. The main components of the charging system are the power module 210, charging interface 220, main controller 230, and power distribution device 240. The power modules 210 are interconnected via DC interconnection buses and control signal lines, forming a distributed power module cluster. The power module 210 converts the AC power input from the power grid into the required DC power and provides the DC power to the charging interface 220. The charging interface 220 connects to the charging gun to output power to the electric vehicle.

[0033] In this scenario example, each power module 210 is equipped with a local controller and a neighbor communication module. The local controller of each power module 210 is used to execute the event-triggered negotiation power control method of the charging device cluster of this application embodiment, specifically including: detecting the heartbeat state of the main controller 230 (i.e., the master control node), and performing mode switching between centralized coordination mode and distributed autonomous mode according to the heartbeat state; performing trigger event detection in the switched operating mode; when no trigger event is detected, maintaining each power module 210 in a zero-negotiation silent state, and maintaining only periodic heartbeat communication between each power module 210; when at least one type of trigger event is detected, initiating distributed negotiation, using the current power relative to the reducible power requirement as a consistency variable to iteratively update each power module 210 until the consistency variable of each power module 210 converges to a globally consistent value; determining the constrained power upper limit of each power module 210 according to the converged globally consistent value; and performing power control according to the constrained power upper limit. The neighbor communication module of each power module 210 is used to realize data interaction between power modules 210, including heartbeat message transmission, consistency variable broadcasting, and neighbor status information reception. The local controller of each power module 210 can be implemented using an STM32 series MCU, embedded processor, or industrial control computer, etc. The neighbor communication module supports communication methods such as CAN bus, RS-485 bus, Ethernet, Wi-Fi, or power line carrier. Each power module 210 is also equipped with a sensor group, including voltage sensors and current sensors, for real-time acquisition of the output voltage and output current of each power module 210. The raw sampled data is filtered by local averaging and then used for status calculation and event judgment.

[0034] Each charging port 220 can be connected to a designated power module 210 via a power distribution device 240, with different charging ports 220 connected to different power modules 210. Each charging port 220 is also connected to a corresponding charging gun for outputting power to the electric vehicle. The charging port 220 and the directly connected power module 210 can be connected via a power switching device or directly with a wire.

[0035] The main controller 230 is connected to each power module 210, each charging interface 220, and the power distribution device 240. The main controller 230 operates in a centralized coordination mode and is used to execute the event-triggered negotiation-based power control method for the charging device cluster according to this embodiment. Specifically, this includes: detecting the heartbeat state of the main controller 230 itself; detecting trigger events in the switched operating mode; maintaining each power module 210 in a zero-negotiation silent state when no trigger event is detected; initiating distributed negotiation when at least one type of trigger event is detected, iteratively updating each power module 210 using the current power relative to the required power reduction ratio as a consistency variable until convergence; determining the constrained power upper limit for each power module 210 based on the converged global consistency value; and executing power control based on the constrained power upper limit, sending the corresponding power upper limit command to each power module 210. The main controller 230 is also used to verify the negotiation consistency of each power module 210 and generate global optimization suggestions. It should be noted that in this embodiment, the main controller 230 is an optional component. When the main controller 230 is running normally, the system works in centralized coordination mode. When the main controller 230 experiences hardware failure, software abnormality, or communication interruption, each power module 210 automatically switches to distributed autonomous mode after detecting continuous heartbeat timeout. It continues to operate by relying only on the local controller and neighbor communication module of each power module 210, and the charging power output is not interrupted during the switching process.

[0036] The power distribution device 240 is connected to each power module 210, each charging interface 220, and the main controller 230. The power distribution device 240 is composed of multiple controllable switches (such as contactors or relays) and is used to control the opening and closing of its internal controllable switches according to scheduling instructions issued by the main controller 230 or the local controllers of each power module 210. This distributes the power output from each power module 210 to each charging interface 220 as needed, and then outputs it to the charging electric vehicle. In this embodiment, the main controller 230 or the local controllers of each power module 210 issue a power upper limit instruction. Each power module 210 adjusts its own output power according to this power upper limit. This power is then aggregated by the power distribution device 240 and output to the target charging interface 220 to meet the total output demand. Through the coordinated control of the power distribution device 240 and each power module 210, it is possible to ensure that each power module 210 outputs power according to its own constrained upper limit, while also flexibly matching the power requirements of different charging interfaces 220.

[0037] In some implementations, the charging system can also be an integrated DC charging pile, with the charging interface 220 connected to the charging gun, the charging gun being hung on the main unit of the charging system via a gun mount on the main body of the charging system, and the power distribution device 240 disposed inside the main body of the charging system.

[0038] Referring to the scenario example of the event-triggered negotiation power control method for charging device clusters in the foregoing embodiments, the event-triggered negotiation power control method for charging device clusters in this application embodiment will be described in detail below.

[0039] Figure 2 This is one of the flowcharts of an event-triggered negotiation power control method for a charging device cluster according to an embodiment of this application. This event-triggered negotiation power control method for the charging device cluster can be executed by the processor of an electronic device, such as... Figure 2 As shown, the event-triggered negotiation power control method for this charging equipment cluster may include the following steps: Step 210: Detect the heartbeat status of the central control node, and switch the operating mode between centralized coordination mode and distributed autonomous mode based on the heartbeat status.

[0040] In this implementation, each device monitors the heartbeat status of the central control node in real time and switches between centralized coordination mode and distributed autonomous mode based on the heartbeat status. Each device determines whether the central control node is in normal working condition by listening to the heartbeat messages periodically sent by the central control node. If it can continuously receive heartbeat messages normally, it maintains or switches to centralized coordination mode; if the heartbeat messages time out continuously, it automatically switches to distributed autonomous mode.

[0041] Optionally, each device has an independent mode state machine. Mode switching is determined independently by each device based on its local heartbeat status detection results, without relying on switching commands from the central control node or other devices. In centralized coordination mode, the central control node is responsible for monitoring the entire site's status, verifying consensus, and generating global optimization suggestions. Each device performs distributed negotiation with the assistance of the central control node. In distributed autonomous mode, each device operates independently, relying only on its local controller and neighbor communication. It performs distributed negotiation based on locally stored neighbor topology information and no longer depends on the central control node. When the central control node resumes normal communication, each device automatically switches back to centralized coordination mode after detecting the heartbeat recovery. The switching process does not interrupt power output. In this embodiment, by detecting the heartbeat status of the central control node and switching between the two modes based on the heartbeat status, the device cluster can utilize the global coordination capabilities of the central control node when it is functioning normally, and can automatically switch to distributed autonomous mode to continue operating when the central control node fails. This eliminates the dependence on a single central control node at the architectural level.

[0042] Step 220: Perform trigger event detection in the switched operating mode.

[0043] In this implementation, each device continuously performs trigger event detection in the switched operating mode to determine whether distributed negotiation needs to be initiated. Trigger event detection is independent of the operating mode; the event detection mechanism operates normally in both centralized coordination mode and distributed autonomous mode. Each device detects whether various trigger events have occurred by monitoring local status information and neighbor communication status in real time.

[0044] Optionally, the triggering events include various types of events that characterize changes in the operating state of the device cluster. The detectors on each device continuously monitor information such as real-time total power, current power of each device, device topology changes, and neighbor heartbeat status to determine if an event has occurred. In this embodiment, by continuously detecting triggering events, an accurate triggering basis is provided for the device cluster to switch from a zero-negotiation silent state to a distributed negotiation state, ensuring that negotiation is only initiated when necessary.

[0045] Step 230: If no triggering event is detected, maintain all devices in a zero-negotiation silent state.

[0046] In this implementation, the device cluster maintains a zero-negotiation silent state when none of the devices detect any triggering events. In this zero-negotiation silent state, devices do not engage in any negotiation-related communication interactions, but only maintain necessary heartbeat communication to detect the online status of neighboring devices.

[0047] Optionally, in zero-negotiation silent mode, each device continues to perform power control according to the existing power limit and does not initiate a new negotiation process. Each device only sends periodic heartbeat messages to neighboring devices at preset intervals to maintain neighbor awareness and online status detection. Negotiation-related data messages are not sent at all when no event is triggered, thereby minimizing communication overhead.

[0048] Optionally, when a previously triggered event exits, the device cluster also enters a zero-negotiation silent state, and each device continues to operate with the current negotiation result without initiating new negotiations. In this embodiment, by maintaining a zero-negotiation silent state when no event is triggered, periodic invalid negotiation communication is fundamentally avoided, significantly reducing the communication bus load and freeing up communication resources.

[0049] Step 240: When at least one type of triggering event is detected, initiate distributed negotiation, using the current power as the proportion of power that can be reduced relative to the required power as the consistency variable to iteratively update each device until the consistency variable of each device converges to a globally consistent value.

[0050] In this implementation, when at least one type of triggering event is detected, the device cluster initiates distributed negotiation. Each device iteratively updates its consistency variable using the ratio of its current power to the required power as the reduction ratio. Through multiple rounds of neighbor communication, the consistency variables of each device gradually converge, eventually reaching a globally consistent value.

[0051] Optionally, each device first calculates an initial consistency variable based on the current actual generated power and the demand power. This variable represents the percentage of power that can be reduced from the current demand power. Each device sends its own consistency variable to its neighboring devices and receives the consistency variables from its neighboring devices. It then uses a preset weight value to weight and merge its own variable and the received neighboring variable to update its local variable. This process is repeated until the deviation between the consistency variables of each device is less than the preset convergence accuracy.

[0052] Optionally, the initial value of the consistency variable reflects the current load reduction capacity of each device. When a device is running at full load, the initial value is zero, indicating that further load reduction is not possible; when a device still has load reduction capacity, the initial value is positive. By using this variable as the convergence target for consistency negotiation, the final converged global consistent value can fairly reflect the actual load reduction capacity of each device, avoiding forcing fully loaded devices to reduce load, thus adapting to the physical constraint that the power demand of each device cannot be increased in one direction. In this embodiment, by using the reduction ratio as the consistency variable for iterative updates, each device can reach a consensus on the load reduction ratio without requiring global information, ensuring that the load reduction obligation is fairly distributed among the devices.

[0053] Step 250: Determine the constrained power limit for each device based on the converged global consistent value.

[0054] In this implementation, once the consistency variables of all devices converge to a globally consistent value, each device calculates its constrained power limit based on this globally consistent value and its own required power. This power limit is simultaneously limited by the physical constraints of each device and the total system capacity constraints, and needs to be verified and adjusted through multi-layered constrained power boundaries.

[0055] Optionally, each device first uses the product of a globally consistent value and its own required power limit as the initial constrained power limit. Then, the validity of this value is verified by the device's own physical constraints and the total system capacity constraints. The value that satisfies all constraints is taken as the device's final power limit.

[0056] Optionally, when the sum of the power limits of all devices exceeds the total system capacity, the globally consistent value is recursively reduced and the power limits of each device are recalculated until the sum of the power limits of all devices does not exceed the total system capacity. In this embodiment, by nested verification of multi-layered constraint power boundaries, it is ensured that the power limits of each device simultaneously meet the individual physical limits of the device and the total system capacity limit, thereby improving the safety and reliability of power control.

[0057] Step 260: Perform power control based on the constrained power upper limit.

[0058] In this embodiment, each device performs power control based on a determined constrained power upper limit. Each device uses this power upper limit as a local power command to adjust its own output power to a level that does not exceed the upper limit.

[0059] Optionally, each device adjusts its output power in real time via a local power drive module, ensuring that the actual output power tracks the constrained power upper limit. During operation, each device continuously collects operating data such as voltage and current, updates its local status information, and uses this updated status information for the next round of event detection and power control, forming a closed-loop control system. In this embodiment, each device performs power control based on the constrained power upper limit, ensuring that the output power of each device always meets multi-layered constraint conditions, and that the entire device cluster operates stably within the total capacity safety boundary.

[0060] In this implementation, by detecting the heartbeat status of the central control node and switching between centralized coordination mode and distributed autonomous mode, it remains silent when there are no events and triggers negotiation when there are events. It uses the power reduction ratio as a consistency variable for iterative convergence, and then determines the constrained power limit of each device based on the converged global consistent value and executes power control. This achieves the technical effect of reducing the periodic invalid communication overhead while avoiding scheduling paralysis caused by a single point of failure of the central control. It solves the technical problems of low reliability and high communication overhead of device clusters due to reliance on centralized central control.

[0061] The steps described above will be explained in detail below.

[0062] Figure 3 This is a second flowchart of an event-triggered negotiation power control method for a charging device cluster according to an embodiment of this application. Figure 3 As shown, step 210, which involves switching between centralized coordination mode and distributed autonomous mode based on the heartbeat status, may include the following steps: Step 211: When the number of consecutive periods without receiving a heartbeat message from the central control node reaches the first preset threshold, switch from centralized coordination mode to distributed autonomous mode; Step 212: When the number of consecutive heartbeat messages received from the central control node in the distributed autonomous mode reaches the second preset threshold, switch back from the distributed autonomous mode to the centralized coordination mode. Step 213: Mode switching is performed independently by each device, and power output is not interrupted during the switching process.

[0063] In this embodiment, when the number of consecutive periods during which each device fails to receive a heartbeat message from the central control node reaches a first preset threshold, it indicates that the central control node may have experienced a hardware failure, software anomaly, or communication interruption. At this time, each device automatically switches from centralized coordination mode to distributed autonomous mode. In distributed autonomous mode, each device operates independently, relying only on its local controller and neighbor communication, and continues to perform event detection, distributed negotiation, and power control, no longer depending on the coordination of the central control node.

[0064] Optionally, when the system is in distributed autonomous mode, each device continuously listens for heartbeat messages from the central control node. When the number of consecutive heartbeat messages received from the central control node in distributed autonomous mode reaches a second preset threshold, it indicates that the central control node has resumed normal communication. Each device automatically switches from distributed autonomous mode back to centralized coordination mode, restoring the central control node's coordination functions such as monitoring the overall station status, verifying consensus, and generating global optimization suggestions. By setting bidirectional switching criteria, switching to distributed autonomous mode when heartbeat timeouts continuously reach a first preset threshold, and switching back to centralized coordination mode when heartbeat recovery continuously reaches a second preset threshold, the device cluster can automatically enter autonomous operation state when the central control fails and automatically return to coordinated operation state when the central control recovers.

[0065] Optionally, mode switching is performed independently by each device, without relying on the central control node or other devices to send switching commands. Each device makes its own switching decision based on its local heartbeat status detection results. During the switching process, each device maintains its current power output without interrupting the charging service. The redistribution of power is completed through a new round of distributed negotiation after the switching is finished, ensuring the continuity of power output. In this embodiment, by having each device independently perform mode switching without interrupting power output, the reliability and service continuity of the mode switching process are guaranteed, achieving seamless connection between the two modes and fully balancing the global optimization capabilities when the central control is normal and the autonomous operation capabilities when the central control fails.

[0066] The events triggered in step 220 can include the following four categories: The total power limit over event is triggered when the real-time total power of all devices exceeds the preset detection threshold of the total capacity limit, or when the real-time total power recovers to the preset exit threshold of the total capacity limit; The single device power deviation event is triggered when the deviation between the current power of any device and the negotiated upper limit of power exceeds a preset deviation threshold. Topology state change events are triggered when a new device is connected, a device is disconnected, or the change in the power demand of a device exceeds a preset threshold. The communication loss event is triggered when the heartbeat messages of a neighboring device time out a preset number of times consecutively.

[0067] In this embodiment, the total power limit exceedance event is used to monitor whether the real-time total power of all devices exceeds the total capacity limit. This event is triggered when the real-time total power exceeds a preset detection threshold of the total capacity limit, activating distributed negotiation to reduce the power of each device. Exit is triggered when the real-time total power recovers to a preset exit threshold of the total capacity limit, restoring each device to its normal negotiated power. By setting detection and exit thresholds near the total capacity limit, frequent event triggering due to instantaneous power fluctuations is avoided.

[0068] Optionally, a single-device power deviation event is used to monitor the deviation between the current power of each device and the negotiated power limit. This event is triggered when the deviation of any device exceeds a preset deviation threshold, initiating a full cluster renegotiation to release excess capacity of that device or adjust power allocation. This event can promptly respond to power deviations caused by changes in device operating conditions (such as reduced power demand), avoiding capacity waste.

[0069] Optionally, topology state change events are used to monitor topology changes in the device cluster, including new device access, existing device removal, or changes in device power demand exceeding a preset threshold. This event is triggered when any of these conditions are met, initiating a full cluster renegotiation to adapt to the power allocation requirements following the topology change. This event can respond promptly to changes in the device cluster's composition, ensuring that the power allocation scheme can be quickly adjusted after a new device is accessed or removed.

[0070] Optionally, the communication loss event is used to monitor the heartbeat message reception status of neighboring devices. When a neighboring device's heartbeat message times out a preset number of times consecutively, the event is triggered, marking the neighboring device as an offline device and triggering the online cluster to reclaim capacity and renegotiate; when the offline device's heartbeat message recovers, it is remarked as an online device and smoothly incorporated into the negotiation cluster.

[0071] In this embodiment, by setting four types of triggering events: total power exceeding limit event, single device power deviation event, topology state change event, and communication disconnection event, the device cluster can fully perceive system state changes and respond in a timely manner. Negotiation is initiated only when an event occurs, which not only ensures the real-time performance of power control, but also avoids the communication overhead caused by periodic invalid communication.

[0072] Figure 4 This is the third flowchart of an event-triggered negotiation power control method for a charging device cluster according to an embodiment of this application. Figure 4As shown, the event-triggered negotiation power control method for the charging equipment cluster may further include the following steps: Step 271: Dynamically adjust the preset detection threshold for total power over-limit events and the preset deviation threshold for single device power deviation events based on the total load rate; Step 272: When the total load rate is higher than the preset high load threshold, reduce the preset detection threshold and the preset deviation threshold; Step 273: When the total load rate is lower than the preset low load threshold, increase the preset detection threshold and the preset deviation threshold.

[0073] In this embodiment, the total load rate is the ratio of the real-time total power of all devices to the upper limit of total capacity. The preset detection threshold for total power exceeding the limit and the preset deviation threshold for single-device power deviation events are dynamically adjusted based on the total load rate. When the total load rate is higher than the preset high load threshold, the detection threshold and deviation threshold are reduced to improve monitoring sensitivity; when the total load rate is lower than the preset low load threshold, the detection threshold and deviation threshold are increased to reduce unnecessary triggering. For example, the preset high load threshold can be set to 80% of the total load rate, and the preset low load threshold can be set to 50% of the total load rate. When the total load rate is higher than 80%, the threshold is reduced to 70% of the base value; when the total load rate is lower than 50%, the threshold is increased to 130% of the base value.

[0074] Optionally, by reducing the threshold at high load rates, the sensitivity of event detection is improved, enabling the device cluster to respond earlier to power limit exceedance risks when operating near full load, effectively preventing overload. Conversely, increasing the threshold at low load rates reduces detection sensitivity, avoiding invalid negotiation triggers caused by minor power fluctuations and further reducing communication overhead. Through these methods, the sensitivity of the trigger threshold is matched to the actual operating state of the device cluster; sensitivity is increased at high loads to prevent overload, and decreased at low loads to reduce unnecessary triggers, achieving intelligent adaptive adjustment of the triggering mechanism.

[0075] Figure 5 This is the fourth flowchart of an event-triggered negotiation power control method for a charging device cluster according to an embodiment of this application. Figure 5 As shown, step 240 initiates distributed negotiation, using the current power as a percentage of the power demand that can be reduced as a consistency variable to iteratively update each device until the consistency variables of each device converge to a globally consistent value. This may include the following steps: Step 241: Send the consistency variables of each device to the neighboring devices, and receive the consistency variables of the neighboring devices; Step 242: Assign weight values ​​to all neighboring devices of each device. The weight values ​​are determined based on the neighbor relationships of the devices in the physical topology. The weight between devices that are not neighbors is zero. Step 243: Based on the weight of each device and the weight of each neighboring device, perform a weighted merging of the consistency variables of each device and the received consistency variables of the neighboring devices, and update the consistency variables of each device. Step 244: Repeat the above sending, receiving and weighted merging process until the deviation between the consistency variables of all devices is less than the preset convergence accuracy.

[0076] In this embodiment, when at least one type of triggering event is detected, the device cluster initiates distributed negotiation. Each device first calculates an initial consistency variable, which represents the proportion of current power that can be reduced relative to the required power, and its value ranges from 0 to 1. Each device sends its own consistency variable to its neighboring devices in its physical topology and receives consistency variables from each neighboring device. Each device assigns weight values ​​to each neighboring device according to the neighbor relationships in the physical topology; the weights between devices that are not neighbors are zero. Each device performs a weighted merging of its own consistency variable and the received neighboring consistency variables based on its own weight and the weights of each neighboring device, and updates its local consistency variable. The above sending, receiving, and weighted merging process is repeated until the deviation between the consistency variables of all devices is less than a preset convergence accuracy.

[0077] Optionally, the initial consistency variable is calculated as follows: when the current actual power output of the equipment is lower than the required power output, the initial consistency variable is the ratio of the difference between the required power output and the actual power output to the required power output; when the current actual power output of the equipment is equal to the required power output, the initial consistency variable is zero. A positive initial consistency variable indicates that the equipment still has room for load reduction under the current demand, while a zero initial consistency variable indicates that the equipment is already running at full load and cannot be further reduced in load.

[0078] Optionally, the weight values ​​are determined as follows: each device assigns a weight to each of its neighboring devices. The weight assigned to a neighbor by a device is inversely proportional to the maximum of the number of neighbors of that device and the number of neighbors of that neighboring device. The weight assigned to a device itself is 1 minus the sum of the weights assigned to all its neighboring devices. This weight determination method allows the weight matrix to be independently calculated by each device based on local neighbor information without requiring global information, while ensuring the symmetry and convergence of the weight matrix.

[0079] In this embodiment, iteration stops when the consistency variables of all devices converge to a globally consistent value. After convergence, each device uses this globally consistent value as a unified load reduction ratio, multiplies it by its own required power, and obtains the constrained power limit, thereby achieving a fair distribution of the load reduction obligation among the devices. This iterative update process only requires communication between each device and its neighboring devices, without requiring global information, and features low communication overhead and strong scalability, making it suitable for real-time control scenarios of large-scale device clusters.

[0080] The weight values ​​in step 242 can be determined as follows: The weight assigned to a neighboring device by the current device is inversely proportional to the maximum of the number of neighbors of the current device and the number of neighbors of that neighboring device. The weight assigned to the current device is 1 minus the sum of the weights assigned to all neighboring devices.

[0081] In this embodiment, each device assigns a weight value to each neighboring device based on the neighbor relationships in the physical topology. The weight assigned by the current device to a particular neighboring device is the reciprocal of the maximum of the number of neighbors of that device and the number of neighbors of that neighboring device. The weight assigned to the current device itself is 1 minus the sum of the weights assigned to all neighboring devices. For devices that are not neighbors, their weights are zero. Through the above determination method, each device only needs to obtain its own and its neighboring devices' neighbor count information to independently calculate its weight, without needing to know the global network topology.

[0082] Optionally, when a device has many neighbors, the weights assigned to each neighboring device are relatively small, thus preventing highly connected devices from having an excessive impact on the overall iteration process. When a device has only a few neighbors, the weights assigned to each neighboring device are relatively large, ensuring that information can be effectively propagated within sparsely connected local areas. The device's own weight is set to 1 minus the sum of all neighbor weights, ensuring that the sum of all weights for each device (including its own weight and all neighbor weights) is 1, guaranteeing the normalization characteristic of the iterative update process.

[0083] Optionally, the above weight determination method allows each device to independently calculate the weight matrix based on its local neighbor information without relying on the central node or global topology information. When the communication topology changes (such as when a new device is added or a device is removed), each device only needs to update its local neighbor information to automatically adapt to the new topology without reconfiguring global parameters. It has good scalability and adaptability and is suitable for large-scale dynamic device cluster control scenarios.

[0084] Figure 6 This is the fifth flowchart of an event-triggered negotiation power control method for a charging device cluster according to an embodiment of this application. Figure 6 As shown, step 250, based on the converged global consensus value, determines the constrained power upper limit for each device, which may include the following steps: Step 251: Calculate the three-level constrained power boundary according to priority nesting. The first priority is the upper limit of the required power of each device, the second priority is the upper limit of the safe power of each device, and the third priority is the upper limit of the total capacity power of all devices. Step 252: Use the product of the globally consistent value and the power requirement limit of each device as the constrained power limit; Step 253: Verify the upper limit of the constrained power using a three-layer constrained power boundary; Step 254: When the total capacity power limit in the three-layer constraints is triggered, calculate the average derating amount that each device can bear under the demand power limit constraint and the safe power limit constraint; Step 255: If the load reduction margin of all equipment is sufficient to bear the equally distributed load reduction, then the equally distributed load reduction shall be performed on each piece of equipment. Step 256: If the load reduction margin of some equipment is insufficient to bear the equally distributed load reduction amount, then the load reduction amount shall be allocated in a weighted manner according to the load reduction margin of each equipment. Step 257: If the total capacity power limit is still exceeded after weighted allocation, overload protection is triggered, and the power of each device is reduced to the preset protection ratio.

[0085] In this embodiment, each device determines its constrained power limit based on the converged global consensus value. First, a three-layer constrained power boundary is calculated using a nested priority structure. The first priority is the power demand limit for each device (i.e., the maximum power currently required by the device), the second priority is the safe power limit for each device (the maximum allowable power determined based on the device's physical safety status), and the third priority is the total capacity power limit for all devices. Then, the product of the global consensus value and the power demand limit for each device is used as the initial constrained power limit, and this initial value is verified using the three-layer constrained power boundary. When the total capacity power limit is triggered, the average load reduction allowance that each device can bear under the constraints of the power demand limit and the safe power limit is calculated. If the load reduction margin of all devices is sufficient to bear the average load reduction allowance, then average load reduction is performed on each device; if the load reduction margin of some devices is insufficient to bear the allowance, then the load reduction allowance is weighted and allocated according to the load reduction margin of each device; if the weighted allocation still exceeds the total capacity power limit, overload protection is triggered, reducing the power of each device to a preset protection ratio.

[0086] Optionally, the priority order of the three-layer constraint power boundaries is fixed and cannot be adjusted. The upper limit of demand power has the highest priority (an inviolable physical hard limit), the upper limit of safety power has the second highest priority (equipment safety fallback), and the upper limit of total capacity power has the lowest priority (system global constraint). When the upper limit of total capacity power is triggered, the system first attempts to perform even load reduction under the constraints of the upper limit of demand and the upper limit of safety. If even load reduction causes some equipment to violate the upper limit of demand or the upper limit of safety, then a weighted allocation is performed according to the load reduction margin of each equipment. If the total power of the system still exceeds the limit after weighted allocation, it indicates that the system is in an extreme overload condition, and overload protection needs to be triggered to forcibly reduce the power of each equipment to a preset protection ratio (such as 50% of the rated power) to ensure system safety.

[0087] Optionally, the load reduction margin of each device corresponds to the consistency variable in step 240, representing the proportion of power that the device can reduce based on its current power demand. Specifically, when the load reduction margin of all devices is sufficient to bear the equally distributed load reduction quota, it means that all devices have enough load reduction space to fairly bear the load reduction obligation; when the load reduction margin of some devices is insufficient to bear the equally distributed load reduction quota, it means that these devices are already operating close to full load or are limited by safety limits and cannot further reduce the load. In this case, their load reduction margin should be exhausted, and the remaining load reduction quota should be allocated to other devices with load reduction space according to their load reduction margin proportion.

[0088] Optionally, through the aforementioned three-layer nested resolution method, the final power limit of each device simultaneously satisfies the demand power limit (physical hard limit), the safe power limit (equipment safety fallback), and the total capacity power limit (system global constraint). Furthermore, the resolution process has a clear priority logic: first, equal distribution; then, weighted allocation; and finally, overload protection, progressing step by step. Compared with existing technologies that constrain the capacity of a single transformer, this application establishes a complete nested constraint system from the individual physical limits of the equipment to the total system capacity, achieving coordinated protection of multi-dimensional safety constraints.

[0089] Figure 7 This is the sixth flowchart of an event-triggered negotiation power control method for a charging device cluster according to an embodiment of this application. Figure 7 As shown, the event-triggered negotiation power control method for the charging equipment cluster may further include the following steps: Step 281: Detect the online status of neighboring devices and mark devices that have not received heartbeat messages for a continuous period of time as offline devices; Step 282: Monitor the instantaneous power consumption of offline devices during actual operation; Step 283: Obtain the estimated power of the offline device, and calculate the release capacity based on the difference between the estimated power and the instantaneous power; Step 284: Add the releasable capacity to the allocable capacity pool and allocate it to each online device according to the power demand ratio of each online device.

[0090] In this embodiment, each device continuously monitors the online status of its neighboring devices. When a device fails to receive a heartbeat message from a neighboring device for an extended period, it marks that neighboring device as offline. For devices marked as offline, the online cluster continuously monitors their actual instantaneous power consumption through electrical detection. The estimated power of the offline device is obtained (e.g., the power value reported last time before going offline or a power value based on historical data statistics). The release capacity is calculated based on the difference between the estimated power and the actual instantaneous power. When the actual instantaneous power is lower than the estimated power, the difference is the release capacity. The calculated release capacity is added to the allocable capacity pool and allocated to each online device according to the proportion of each online device's required power to the total required power of all online devices, thus updating the available power limit of each online device.

[0091] Optionally, when a device is marked as offline, it continues to operate locally in response to the actual needs of the vehicle's BMS, without stopping the charging service due to communication interruption. Online devices do not conservatively reserve capacity for offline devices; instead, they dynamically calculate and recycle the release capacity based on the actual instantaneous power of the offline device, thereby avoiding capacity waste caused by conservative reservations.

[0092] Optionally, once an offline device resumes communication, the online cluster re-marks it as an online device and incorporates its current power and status information into subsequent distributed negotiations, enabling a smooth rejoin. This approach breaks the existing technical bias that offline devices must conservatively reserve capacity, achieving dynamic reclamation of idle capacity from offline devices and on-demand allocation of capacity to online devices, effectively improving the overall capacity utilization of the system.

[0093] Figure 8 This is flowchart seven of an event-triggered negotiation power control method for a charging device cluster according to an embodiment of this application. Figure 8 As shown, the event-triggered negotiation power control method for the charging equipment cluster may further include the following steps: Step 291: Check if the device cluster is not in a fully configured state; Step 292: When the device is not fully configured, determine the idle capacity value based on the relationship between the total capacity power limit, the sum of the current actual power of all online devices, and the sum of the current actual power of all offline devices; Step 293: When the idle capacity value is greater than zero, allocate the idle capacity to each online device according to the proportion of the power demand of each online device to the total power demand of all online devices.

[0094] In this embodiment, each device detects whether the device cluster is in a non-fully utilized state. A non-fully utilized state refers to a situation where the total actual operating power of the online and offline devices connected to the device cluster has not yet reached the total capacity power limit. When the device cluster is in a non-fully utilized state, the idle capacity value is determined based on the difference between the total capacity power limit and the sum of the current actual power of all online devices and the sum of the current actual power of all offline devices. When this idle capacity value is greater than zero, it indicates that there is unused capacity in the device cluster. In this case, the idle capacity is allocated to each online device according to the proportion of the demand power of each online device to the total demand power of all online devices, and the available power limit of each online device is updated.

[0095] Optionally, the idle capacity value can be determined as follows: calculate the total capacity power limit, subtract the sum of the current actual power of all online devices, and then subtract the sum of the current actual power of all offline devices. When the calculation result is greater than zero, the difference is the idle capacity value. If the calculation result is less than or equal to zero, it indicates that the device cluster is in a fully configured or overloaded state, and there is no idle capacity available for allocation.

[0096] Optionally, when the idle capacity value is greater than zero, each online device, while meeting its own required power limit and safe power limit, receives additional power allocation according to the demand ratio, and can temporarily operate at a level higher than the rated power. This allocation method makes full use of the unused capacity of the transformer in a multi-pile shared scenario, avoiding capacity waste caused by conservative configuration.

[0097] Optionally, once idle capacity is allocated to online devices, a capacity reallocation process is triggered if a new device is connected. Specifically, when the system detects a new device connection, it executes an appropriate allocation or reclamation strategy based on the relationship between the currently available idle capacity and the power demand of the new device. If there is sufficient idle capacity, it is allocated directly; if there is insufficient idle capacity, capacity is proportionally reclaimed from online devices that have already received excess power and transferred to the new device, achieving a smooth transition when a new device connects. Through this method, the device cluster can fully utilize the remaining transformer capacity even when not fully utilized, improving the overall capacity utilization of the system and the charging speed of online devices.

[0098] Based on the above embodiments, this application also provides a charging system, including at least two power modules, at least one charging interface, a power distribution device, and a main controller; the main controller is connected to each power module, each charging interface, and the power distribution device respectively, and the main controller is used to execute the event-triggered negotiation power control method of the charging device cluster as described above; the power distribution device is connected to each power module and each charging interface respectively.

[0099] It should be noted that for details not disclosed in the charging system of this embodiment, please refer to the details disclosed in the embodiment of the event-triggered negotiation power control method for charging device cluster in this specification, which will not be repeated here.

[0100] Based on the above embodiments, the present invention also provides a non-transitory computer-readable storage medium storing a computer program thereon. When executed by a processor, the computer program implements an event-triggered negotiation power control method for a charging device cluster provided by the above methods. The method includes: detecting the heartbeat state of the master control node; switching the operating mode between a centralized coordination mode and a distributed autonomous mode based on the heartbeat state; and detecting trigger events in the switched operating mode. When no triggering event is detected, all devices remain in a zero-negotiation silent state; when at least one type of triggering event is detected, distributed negotiation is initiated, and each device is iteratively updated with the current power as a percentage of the power required to be reduced as a consistency variable until the consistency variable of each device converges to a globally consistent value; based on the converged globally consistent value, the constrained power limit of each device is determined. Power control is performed based on the constrained power upper limit.

[0101] Optionally, such as Figure 9 As shown, the main controller may include a processor 910, a communication interface 920, a memory 930, and a communication bus 940. The processor 910, communication interface 920, and memory 930 communicate with each other via the communication bus 940. The processor 910 can call logic instructions in the memory 930 to execute an event-triggered negotiation power control method for the charging device cluster. This method includes: acquiring the operating parameters of each target module, including the electrical and thermodynamic parameters of each target module; analyzing the efficiency curve parameters of each target module based on the operating parameters; predicting the future trend of efficiency curve parameter changes for each target module based on historical efficiency curve parameters and evaluating the confidence level of the prediction results; determining the target control mode corresponding to each target module based on the confidence level; and under the target control mode, constructing a multi-objective optimization model with the objectives of maximizing efficiency and balancing lifetime based on the efficiency curve parameter change trends, and solving for the optimal operating parameter allocation scheme for each target module through the multi-objective optimization model.

[0102] Furthermore, the logical instructions in the aforementioned memory 930 can be implemented as software functional units and, when sold or used as independent products, can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, essentially, or the part that contributes to the prior art, or a part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods of the various embodiments of the present invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0103] The above are merely specific embodiments of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

[0104] The above are merely preferred embodiments of this application and are not intended to limit this application. Various modifications and variations can be made to this application by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the protection scope of this application.

Claims

1. An event-triggered negotiation-based power control method for a charging equipment cluster, characterized in that, The method is applied to a charging system, which includes at least two power modules and at least one charging interface, wherein the power modules are connected to each other via a controllable switch; the method includes: Detect the heartbeat status of the central control node, and switch the operating mode between centralized coordination mode and distributed autonomous mode based on the heartbeat status; Trigger event detection is performed in the switched operating mode; When no triggering event is detected, all devices remain in a zero-negotiation silent state; When at least one type of triggering event is detected, distributed negotiation is initiated, and each device is iteratively updated using the current power as a percentage of the power required to be reduced as a consistency variable until the consistency variable of each device converges to a globally consistent value. Based on the converged global consensus value, determine the constrained power limit for each device; Power control is performed based on the power upper limit after the constraints are applied.

2. The event-triggered negotiation power control method for a charging equipment cluster according to claim 1, characterized in that, The step of switching the operating mode between centralized coordination mode and distributed autonomous mode based on the heartbeat status includes: When the number of consecutive periods without receiving a heartbeat message from the central control node reaches a first preset threshold, the system switches from the centralized coordination mode to the distributed autonomous mode. When the number of consecutive heartbeat messages received from the central control node in the distributed autonomous mode reaches a second preset threshold, the system switches back from the distributed autonomous mode to the centralized coordination mode. The mode switching is performed independently by each device, and the power output is not interrupted during the switching process.

3. The event-triggered negotiation power control method for a charging equipment cluster according to claim 1, characterized in that, The triggering events include the following four categories: The total power limit over event is triggered when the real-time total power of all devices exceeds the preset detection threshold of the total capacity limit, or when the real-time total power recovers to the preset exit threshold of the total capacity limit; The single device power deviation event is triggered when the deviation between the current power of any device and the negotiated upper limit of power exceeds a preset deviation threshold. Topology state change events are triggered when a new device is connected, a device is disconnected, or the change in the power demand of a device exceeds a preset threshold. The communication loss event is triggered when the heartbeat messages of a neighboring device time out a preset number of times consecutively.

4. The event-triggered negotiation power control method for a charging equipment cluster according to claim 3, characterized in that, The method further includes: The preset detection threshold for the total power over-limit event and the preset deviation threshold for the single device power deviation event are dynamically adjusted based on the total load rate. When the total load rate is higher than the preset high load threshold, the preset detection threshold and the preset deviation threshold are reduced. When the total load rate is lower than the preset low load threshold, the preset detection threshold and the preset deviation threshold are increased.

5. The event-triggered negotiation power control method for a charging equipment cluster according to claim 1, characterized in that, The process of initiating distributed negotiation involves iteratively updating each device using the ratio of the current power to the required power reduction as a consistency variable, until the consistency variable of each device converges to a globally consistent value. This includes: Each device sends its consistency variables to neighboring devices and receives the consistency variables from the neighboring devices. Assign weight values ​​to all neighboring devices of each device, wherein the weight values ​​are determined based on the neighbor relationships of the devices in the physical topology, and the weight between devices that are not neighbors is zero; Based on the weight of each device and the weight of each neighboring device, the consistency variables of each device and the received consistency variables of the neighboring devices are weighted and merged, and the consistency variables of each device are updated. Repeat the above sending, receiving, and weighted merging process until the deviation between the consistency variables of all devices is less than the preset convergence accuracy.

6. The event-triggered negotiation power control method for a charging equipment cluster according to claim 5, characterized in that, The weight values ​​are determined as follows: The weight assigned to the neighboring device by the current device is inversely proportional to the maximum of the number of neighbors of the current device and the number of neighbors of the neighboring device. The weight assigned to the current device is 1 minus the sum of the weights assigned to all the neighboring devices.

7. The event-triggered negotiation power control method for a charging equipment cluster according to claim 1, characterized in that, The step of determining the constrained power upper limit for each device based on the converged global consensus value includes: The three-level constraint power boundary is calculated in a nested manner according to priority. The first priority is the upper limit of the required power of each device, the second priority is the upper limit of the safe power of each device, and the third priority is the upper limit of the total capacity power of all devices. The product of the globally consistent value and the power requirement limit of each device is used as the constrained power limit; The upper limit of the constrained power is verified by the three-layer constrained power boundary. When the total capacity power limit in the three-layer constraint is triggered, calculate the average load reduction amount that each device can bear under the demand power limit constraint and the safe power limit constraint. If the load reduction margin of all equipment is sufficient to bear the average load reduction amount, then the average load reduction shall be performed on each piece of equipment. If the load reduction margin of some equipment is insufficient to bear the equally distributed load reduction amount, the load reduction amount shall be allocated in a weighted manner according to the load reduction margin of each equipment. If the total power capacity is still exceeded after weighted allocation, overload protection will be triggered, reducing the power of each device to the preset protection ratio.

8. The event-triggered negotiation power control method for a charging equipment cluster according to claim 1, characterized in that, The method further includes: Detect the online status of neighboring devices and mark devices that have not received heartbeat messages for a continuous period of time as offline devices; Monitor the instantaneous power of the offline device during actual operation; Obtain the estimated power of the offline device, and calculate the release capacity based on the difference between the estimated power and the instantaneous power; The releasable capacity is added to the allocable capacity pool and allocated to each online device according to the power demand ratio of each online device.

9. The event-triggered negotiation power control method for a charging equipment cluster according to claim 1, characterized in that, The method further includes: Is the detection equipment cluster not in a fully configured state? When the device is not fully equipped, the idle capacity value is determined based on the relationship between the total capacity power limit, the sum of the current actual power of all online devices, and the sum of the current actual power of all offline devices. When the idle capacity value is greater than zero, the idle capacity is allocated to each of the online devices according to the proportion of the power demand of each of the online devices to the total power demand of all the online devices.

10. A charging system, characterized in that, It includes at least two power modules, at least one charging interface, a power distribution device, and a main controller; The main controller is connected to each of the power modules, each of the charging interfaces and the power distribution device respectively, and the main controller is used to execute the event-triggered negotiation power control method of the charging equipment cluster as described in any one of claims 1 to 9; The power distribution device is connected to each of the power modules and each of the charging interfaces respectively.

Images