Micro-grid edge side ai collaborative optimization control method and system

Abstract

The application discloses a kind of microgrid edge side AI collaborative optimization control method and system, it is related to microgrid control technical field, including: according to edge device operating state, device dynamic response model and the feedback correction of last control cycle, the active power regulation range of equipment is calculated;The total active power curve of microgrid issued is decomposed to the range, and the active power instruction sequence of equipment is formed;According to real-time power flow calculation result and power sensitivity, electrical constraint correction is carried out, and the corrected instruction sequence and power correction are formed;With power correction quantity, equipment power tracking deviation is constrained, and the feedback correction of next control cycle is generated and rewritten;Switch to local autonomous control when communication is abnormal, and state synchronization is carried out after recovery.Can total active power target be converted into executable instruction that meets device dynamic response capability, operating boundary constraint and electrical operation limit, reduce cross-cycle deviation accumulation, improve edge side control continuity and security.

Description

Technical Field

[0001] This invention relates to the field of microgrid control technology, specifically to a microgrid edge-side AI collaborative optimization control method and system. Background Technology

[0002] With the continuous advancement of the construction of new power systems, the power regulation tasks undertaken by microgrids are becoming increasingly heavy, and the scale of adjustable resources connected to microgrids is also growing rapidly. In order to improve the overall operation level, microgrids generally adopt a hierarchical control architecture, that is, the server side first performs global optimization decisions to generate power control targets for the entire microgrid, and then sends the targets down to the edge side, where they are converted into control commands for specific devices and executed.

[0003] In related technologies, the edge side typically allocates the total target directly to each access device according to predetermined rules. This allocation process lacks a complete characterization of the actual executable capabilities of the devices, and the resulting control commands can easily exceed the current adjustment levels that the devices can handle. The allocation results often lack network-level electrical verification before being issued, which may cause local electrical quantities to exceed permissible limits after command execution. Deviations between device execution results and control targets are also difficult to attribute reasonably and accumulate gradually during continuous control, thus deteriorating the tracking effect of the total target. Once communication between the edge side and the server side becomes abnormal, the edge side will struggle to maintain continuous control due to the lack of reliable local control data.

[0004] Therefore, how to convert total control targets into control results that are executable by equipment and meet electrical operation requirements at the edge of a microgrid has become a technical problem that urgently needs to be solved in this field. Summary of the Invention

[0005] The purpose of this invention is to provide a microgrid edge-side AI collaborative optimization control method and system to solve the problems mentioned in the background art.

[0006] To achieve the above objectives, the present invention provides the following technical solution:

[0007] In a first aspect, the present invention provides an AI collaborative optimization control system for the edge side of a microgrid, comprising:

[0008] The active power adjustment range calculation module is used to calculate the active power adjustment range of the equipment based on the operating status of the edge equipment, the dynamic response model of the equipment, and the feedback correction amount of the previous control cycle.

[0009] The strategy rolling decomposition module is used to decompose the total active power curve of the microgrid issued by the AI ​​server into the active power adjustment range of the device, forming a sequence of active power command for the device.

[0010] The electrical constraint correction module is used to perform electrical constraint correction on the active power command sequence of the equipment according to the real-time power flow calculation results and power sensitivity, so as to form the corrected active power command sequence of the equipment and the power correction amount;

[0011] The closed-loop feedback module is used to constrain the power tracking deviation of the equipment with the power correction amount, generate the feedback correction amount for the next control cycle, and use the feedback correction amount for the next control cycle as the input of the active power adjustment range calculation module of the equipment.

[0012] The local autonomous control module is used to switch the edge-side control basis according to the communication status with the AI ​​server, and to obtain local control commands and synchronous operation status.

[0013] Secondly, this invention provides a microgrid edge-side AI collaborative optimization control method, implemented based on the aforementioned system, comprising:

[0014] The active power adjustment range of the equipment is calculated based on the operating status of the edge device, the dynamic response model of the device, and the feedback correction amount of the previous control cycle.

[0015] The total active power curve of the microgrid issued by the AI ​​server is decomposed into the active power adjustment range of the equipment to form a sequence of active power commands for the equipment.

[0016] The active power command sequence of the equipment is electrically constrained and corrected based on the real-time power flow calculation results and power sensitivity, resulting in the corrected active power command sequence and power correction amount.

[0017] The power correction amount constrains the power tracking deviation of the equipment, generates the feedback correction amount for the next control cycle, and uses the feedback correction amount for the next control cycle as the input for calculating the active power adjustment range of the equipment in the next calculation.

[0018] When the communication status with the AI ​​server is abnormal, the most recently formed corrected device active power command sequence is used as the basis for local control. After the communication status with the AI ​​server is restored, status synchronization is performed to obtain local control commands and synchronized operating status.

[0019] The technical effects and advantages provided by the present invention in the above technical solution are as follows:

[0020] This invention calculates the active power adjustment range of equipment at the microgrid edge based on the equipment's dynamic response model and the feedback correction amount from the previous control cycle. The total active power curve issued by the AI ​​server is then decomposed into this range, ensuring that the active power command sequence matches the equipment's executable capacity. This avoids the direct allocation of total targets, preventing out-of-bounds commands. Before command issuance, the command sequence is electrically constrained based on real-time power flow calculations and power sensitivity. The decomposed result is retained with the smallest possible power correction amount while meeting electrical operating limits, reducing the risk of electrical quantity exceeding limits during command execution. After command execution, the power correction amount constrains the equipment's power tracking deviation. The remaining deviation after deducting planned deviations is converted into the feedback correction amount for the next control cycle and written back, achieving reasonable deviation attribution and cross-cycle elimination, suppressing deviation accumulation in continuous control. When communication with the AI ​​server is abnormal, the most recently electrically constrained command sequence is used as the basis for local control, and the state is synchronized after communication is restored, ensuring continuous operation of edge-side control. Attached Figure Description

[0021] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in this invention. For those skilled in the art, other drawings can be obtained based on these drawings.

[0022] Figure 1 A schematic diagram of the architecture of a microgrid edge-side AI collaborative optimization control system provided in an embodiment of the present invention;

[0023] Figure 2 A schematic diagram illustrating the process of determining the active power adjustment range of the device according to an embodiment of the present invention;

[0024] Figure 3 This is a schematic diagram of the processing procedure of the closed-loop feedback module provided in an embodiment of the present invention;

[0025] Figure 4 This is a flowchart illustrating a microgrid edge-side AI collaborative optimization control method provided in an embodiment of the present invention.

[0026] In the diagram: 100, Microgrid edge control terminal; 101, Equipment active power adjustment range calculation module; 102, Strategy rolling decomposition module; 103, Electrical constraint correction module; 104, Closed-loop feedback module; 104-1, Deviation calculation unit; 104-2, Feedback correction amount calculation unit; 104-3, Write-back control unit; 105, Local autonomous control module; 200, AI server; 300, Grid-connected measurement and control device; 400, Data acquisition device; 500, Edge device controller; 600, Edge device. Detailed Implementation

[0027] Exemplary embodiments will now be described more fully with reference to the accompanying drawings. However, these exemplary embodiments can be implemented in many forms and should not be construed as limited to the examples set forth herein; rather, they are provided to make the description of this application more complete and comprehensive, and to fully convey the concept of the exemplary embodiments to those skilled in the art. The drawings are merely illustrative illustrations of this application and are not necessarily drawn to scale. The same reference numerals in the drawings denote the same or similar parts, and therefore repeated descriptions of them will be omitted.

[0028] Furthermore, the described features, structures, or characteristics can be combined in any suitable manner in one or more exemplary embodiments. Numerous specific details are provided in the following description to give a full understanding of the exemplary embodiments disclosed in this application. However, those skilled in the art will recognize that the technical solutions disclosed in this application can be practiced with one or more specific details omitted, or other methods, components, steps, etc., can be employed. In other instances, well-known structures, methods, implementations, or operations are not shown or described in detail to avoid obscuring various aspects of the disclosure of this application.

[0029] Example 1

[0030] like Figure 1 As shown, this embodiment discloses a microgrid edge-side AI collaborative optimization control system. The system is installed at the microgrid edge-side control terminal 100 and is used to convert the total active power curve of the microgrid issued by the AI ​​server 200 into equipment control results that meet the dynamic response capability, operating boundary constraints, and electrical operating limits of the equipment at the edge side. The system includes:

[0031] The active power adjustment range calculation module 101 is used to calculate the active power adjustment range of the equipment based on the operating status of the edge equipment, the dynamic response model of the equipment, and the feedback correction amount of the previous control cycle.

[0032] The strategy rolling decomposition module 102 is used to decompose the total active power curve of the microgrid issued by the AI ​​server into the active power adjustment range of the device, forming a sequence of active power instructions for the device.

[0033] The electrical constraint correction module 103 is used to perform electrical constraint correction on the active power command sequence of the equipment according to the real-time power flow calculation results and power sensitivity, so as to form the corrected active power command sequence of the equipment and the power correction amount.

[0034] The closed-loop feedback module 104 is used to constrain the power tracking deviation of the equipment with the power correction amount, generate the feedback correction amount for the next control cycle, and use the feedback correction amount for the next control cycle as the input of the active power adjustment range calculation module 101 of the equipment.

[0035] The local autonomous control module 105 is used to switch the edge-side control basis according to the communication status with the AI ​​server, and to obtain local control instructions and synchronous operation status.

[0036] The microgrid edge control terminal 100 is communicatively connected to the AI ​​server 200, and also to the edge device controller 500, grid-connected monitoring and control device 300, and data acquisition device 400 within the microgrid. The AI ​​server 200 sends the total active power curve of the microgrid to the edge control terminal 100, which represents the total active power target of the microgrid within a continuous control cycle. The edge device controller 500 uploads edge device operating measurements, which represent the operating status of the edge device at the corresponding sampling time. The grid-connected monitoring and control device 300 provides operating data corresponding to the grid connection point active power, node voltage, branch power, and electrical operating limits. The data acquisition device 400 provides edge device status-related data.

[0037] It is understood that the grid-connected monitoring and control device 300, data acquisition device 400, and edge device controller 500 can be devices independently of the microgrid edge-side control terminal 100, communicating with the microgrid edge-side control terminal 100 via a communication interface; or they can be integrated into the microgrid edge-side control terminal 100 as functional modules, in which case the aforementioned communication connection is achieved through internal data interaction methods such as bus transmission, inter-process communication, or shared storage within the microgrid edge-side control terminal 100; or they can be a combined deployment where some devices are independently set up and the rest are integrated into the microgrid edge-side control terminal 100. This embodiment does not limit the deployment form of the grid-connected monitoring and control device 300, data acquisition device 400, and edge device controller 500. Figure 1 The deployment configuration shown is only one example, and none of the above deployment configurations affect the processing of each module in this embodiment.

[0038] The edge devices 600 are access devices within the microgrid with active power regulation capabilities, including energy storage converters, photovoltaic inverters, adjustable load control devices, and charging pile control devices. The operational measurements of the edge devices include the current measured active power and device status parameters: the energy storage converter's status parameters include state of charge, charging / discharging status, and available capacity; the photovoltaic inverter's status parameters include current output, available output, and power limiting status; the adjustable load control device's status parameters include current load power and adjustable status; and the charging pile control device's status parameters include current charging power and charging execution status. This data is uploaded by the corresponding edge device controller 500 to the microgrid edge-side control terminal 100 according to the sampling timescale, serving as the data basis for subsequent processing by various modules.

[0039] It should be noted that the total active power curve of the microgrid is target data at the overall microgrid level and cannot directly represent the executable power of a single edge device within the current control cycle. The active power regulation that an edge device can undertake is jointly limited by the current measured active power, device status variables, active power ramp-up limits, response delay, rated active power range, and allowable range of status variables. Therefore, the microgrid edge-side control terminal 100 first calculates the active power regulation range of the devices, then performs strategy rolling decomposition within this range, and performs electrical constraint correction before issuing and executing commands.

[0040] In the active power adjustment range calculation module 101, the microgrid edge control terminal 100 forms the edge device operating status based on the edge device operating quantity measurement value, which is used to characterize the actual operating conditions of the edge device in the current control cycle; the feedback correction amount of the previous control cycle is provided by the closed-loop feedback module 104, which represents the remaining tracking deviation that needs to be processed in the subsequent control cycle after deducting the planning deviation in the previous control cycle.

[0041] The device dynamic response model includes active power response relationship, operating boundary constraints, and energy conversion efficiency parameters. The active power response relationship is characterized by active power ramp-up limit and response delay. The operating boundary constraints are characterized by rated active power range and allowable range of state variables. The energy conversion efficiency parameters are used to constrain the active power allocation of the device when the total active power curve of the microgrid is decomposed to the edge device.

[0042] The device dynamic response model is pre-configured on the microgrid edge-side control terminal 100 based on the nameplate parameters of the edge devices, the technical parameters of the device controllers, and historical operating data, and corresponds to the edge device identifier. The active power ramp-up limit indicates the allowable increase or decrease in active power of the edge device within a control cycle; response delay indicates the time required for the edge device to form an effective active power change after receiving a control command; rated active power range indicates the active power setting range that the edge device controller 500 is allowed to receive; and the state quantity allowable range indicates the operating state range that the edge device must still meet after performing active power regulation. Specifically, the state quantity allowable range for energy storage converters includes the state of charge allowable range and charge / discharge allowable conditions; the state quantity allowable range for photovoltaic inverters includes available output conditions and limited output conditions; and the state quantity allowable range for charging pile control devices includes charging execution status and allowable load reduction conditions. The energy conversion efficiency parameter reflects the differences in energy conversion when different edge devices undertake the same active power regulation.

[0043] As a feasible scenario, the total active power curve of a microgrid requires a reduction in the power purchased by the grid connection point during the current control period: the rated discharge power of some energy storage converters meets the allocation requirements, but their state of charge (SOC) is close to the lower limit of the allowable SOC range; the SOC of some energy storage converters meets the discharge requirements, but is constrained by active power ramp-up limitations and response delays; the charging pile control device has load reduction capabilities, but the charging execution status limits the load reduction range. Therefore, the total active power curve of the microgrid should first be constrained by the equipment dynamic response model to limit the active power adjustment range of the equipment, so as to avoid edge equipment from exceeding the operating boundary or generating unexecutable commands due to bearing the total target.

[0044] Specifically, the step of calculating the active power adjustment range of the device based on the operating status of the edge device, the device dynamic response model, and the feedback correction amount of the previous control cycle includes:

[0045] The operational measurement values ​​of edge devices are time-aligned according to the control cycle to obtain the cycle operation status;

[0046] The sampling time, communication return time, and edge-side reception time of each edge device differ. The microgrid edge-side control terminal 100 uses the control cycle as a unified time reference. Based on the sampling timestamp carried by the measured values, it merges the current measured active power and device status quantities belonging to the same control cycle into the same cycle's data set: for measured values ​​with a sampling frequency higher than the control cycle, it selects the valid measured values ​​corresponding to the control time within the control cycle; for measured values ​​with a sampling frequency lower than the control cycle but within the valid hold time, it maps them to the corresponding control cycle according to the sampling timestamp; for measured values ​​that are updated multiple times within the same control cycle, it uses data whose sampling timestamp conforms to the current control cycle. Through the above processing, the edge devices' data within the control cycle are obtained. The periodic operation status, in which the current measured active power and the equipment status are in the same control period, can ensure the consistency of the data on which the calculation of the equipment's active power adjustment range is based.

[0047] The initial active power adjustment range of the equipment is determined based on the cycle operation status and the equipment dynamic response model;

[0048] like Figure 2 As shown, the active power adjustment range calculation module 101 reads the corresponding dynamic response model of the edge device based on the edge device identifier, calculates the state-allowed power range and ramp constraint power range for each edge device, and finds their intersection to obtain the initial active power adjustment range of the device. The initial active power adjustment range of the device represents the range of active power that the edge device is allowed to output or absorb within the current control cycle based on the device state boundary and dynamic response capability, without adding the feedback correction amount from the previous control cycle. Calculating the initial active power adjustment range of the device first, and then introducing the feedback correction amount, can distinguish between the device's own executable capability and the cross-cycle feedback compensation requirement.

[0049] The step of determining the initial active power adjustment range of the equipment based on the periodic operating state and the equipment dynamic response model includes:

[0050] Based on the equipment state variables and operating boundary constraints in the cycle operation state, the upper and lower limits of the active power that can be output or absorbed in the current control cycle are determined, and the state-allowed power range is obtained.

[0051] The active power adjustment range calculation module 101 determines the upper and lower limits of the active power that the edge device can output or absorb in the current control cycle based on the position of the device's state variables within the allowable range of state variables. This, together with the rated active power range, forms the state-allowed power range: For energy storage converters, the upper and lower limits of the active power allowed for charging or discharging are determined based on the current state of charge, the allowable state of charge range, the control cycle length, and the rated active power range; for photovoltaic inverters, the upper and lower limits of the active power allowed for output are determined based on available output and the power limiting state; for charging pile control devices, the upper and lower limits of the active power allowed for load reduction or recovery are determined based on the current charging power and the charging execution state. The state-allowed power range prevents edge devices from exceeding the operating boundary due to assuming the active power allocation of the equipment.

[0052] As one possible implementation, taking the active power output from the edge device to the microgrid as the positive direction, for the energy storage converter, the upper and lower limits of the permissible power range under the current state are determined by the following formula:

[0053]

[0054]

[0055] in, For edge devices During the control cycle The state of charge, and These are the upper and lower limits of the permissible range of the state of charge, respectively. The rated energy storage capacity is expressed in kilowatt-hours. The control cycle duration is measured in hours. and These are the rated discharge power and the rated charge power, respectively. and These are the discharge efficiency and the charging efficiency, respectively, both of which are determined by the energy conversion efficiency parameter.

[0056] For photovoltaic inverters, the permissible power range under certain conditions is determined by the following formula based on available output and power curtailment status:

[0057]

[0058] For adjustable load control devices and charging pile control devices, the permissible power range for each state is determined according to the following formula, based on the adjustable state and the charging execution state respectively:

[0059]

[0060]

[0061] in, For the current available output of the photovoltaic inverter, The upper limit of the allowable output power corresponding to the limited power generation state is taken as the upper limit of the rated active power range when there is unlimited power generation. The output power of the photovoltaic inverter can be reduced to zero. and These are the rated load power of the adjustable load control device and the minimum guaranteed load power determined by the adjustable state, respectively. and These are the rated charging power of the charging pile control device and the minimum sustaining charging power determined by the charging execution status, respectively. The negative sign indicates that the two types of devices operate by absorbing active power.

[0062] Based on the current measured active power and active power response relationship in the cycle operation state, the upper and lower limits of active power that can be reached in the next control cycle are determined, and the ramp constraint power range is obtained.

[0063] The active power adjustment range calculation module 101 uses the current measured active power in the cyclic operation state as a benchmark, and determines the upper limit of the allowable power increment and the lower limit of the power reduction within the control cycle according to the active power ramp-up limit. When there is a response delay, it corrects the time for effective power change according to the control cycle time occupied by the response delay, ensuring that the upper and lower limits of active power achievable in the next control cycle are consistent with the actual response process of the equipment. The ramp-up constraint power range reflects the dynamic reachability of the equipment and can reduce the deviation between the equipment's active power command sequence and the actual executed power.

[0064] As one possible implementation method, the upper and lower limits of the climbing constraint power range are determined by the following formula:

[0065]

[0066] in, For edge devices During the control cycle The current measured active power, and These are the power rise rate limit and power fall rate limit determined by the active power ramp limit, respectively. Let the response delay be denoted as . and The larger of the difference and zero is taken as the time during which an effective power change can be formed within the control cycle; when the response delay is not less than the control cycle duration, the ramp constraint power range converges to the current measured active power, that is, no effective power change is formed within the control cycle.

[0067] After mapping the state-allowed power range and the ramp-constrained power range to the same control time period, the intersection is calculated, and the initial equipment active power adjustment range is obtained by the following formula:

[0068]

[0069] Where, in the formula For edge devices During the control cycle The initial active power adjustment range of the equipment. and These are the lower and upper limits of the allowable power range for the stated state, respectively. and These are the lower and upper limits of the climbing constraint power range, respectively.

[0070] When finding the intersection, the permissible power range and the ramp constraint power range both correspond to the same control cycle. and The maximum value in the range is taken as the lower limit of the range. and The minimum value in the interval is used as the upper limit of the range. This intersection operation makes the initial active power adjustment range of the equipment subject to both equipment state variables and dynamic response capabilities, which can improve the feasibility of subsequent active power allocation.

[0071] The feedback correction from the previous control cycle is applied to the initial active power adjustment range of the equipment, and the correction result is restricted within the operating boundary constraints to obtain the active power adjustment range of the equipment.

[0072] The feedback correction from the previous control cycle serves as the processing basis for the active power adjustment range calculation module 101, rather than being directly used as the control command for the edge device. If the actual output power of a certain edge device in the previous control cycle is lower than the command power corresponding to the corrected active power command sequence, the active power adjustment range calculation module 101 adjusts the initial active power adjustment range of the edge device according to the feedback correction generated by the closed-loop feedback module 104: if the adjusted range is within the rated active power range and the allowable range of the state quantity, it is used as the active power adjustment range; if it exceeds the operating boundary constraints, the active power adjustment range is obtained after being limited by the operating boundary constraints.

[0073] As one possible implementation, the feedback correction from the previous control cycle is applied to the initial active power adjustment range of the equipment by shifting the interval endpoints, and is truncated according to the rated active power range, i.e.:

[0074]

[0075]

[0076] in, and These are the lower and upper limits of the equipment's active power adjustment range, respectively. and These are the lower and upper limits of the active power adjustment range of the initial equipment, respectively. To allocate to edge devices During the control cycle The feedback correction component of the previous control cycle. and These are the lower and upper limits of the rated active power range, respectively; the power boundary corresponding to the allowable range of state variables is already reflected by the allowable power range of state variables.

[0077] It should be noted that introducing the feedback correction from the previous control cycle does not mean forcing the edge devices to complete all remaining tracking deviations in the current control cycle. If the operating boundary constraints do not allow for the complete absorption of this feedback correction, then the active power adjustment range of the devices will only be formed within the range allowed by the operating boundary constraints, thereby achieving cross-cycle feedback correction and preventing edge devices from operating beyond their limits. After each edge device obtains its corresponding active power adjustment range, the active power adjustment range calculation module 101 outputs it to the strategy rolling decomposition module 102 to prevent the total active power curve of the microgrid from being decomposed into instructions that exceed the executable capabilities of the edge devices.

[0078] In the strategy rolling decomposition module 102, the microgrid edge-side control terminal 100 receives the total active power curve of the microgrid from the AI ​​server 200, reads the active power adjustment range of the equipment and the feedback correction amount of the previous control cycle, and outputs the active power command sequence of the equipment. The active power command sequence of the equipment includes the control cycle. The active power allocation corresponding to each edge device participating in the allocation, and each active power allocation is within the active power adjustment range of the corresponding edge device.

[0079] Furthermore, the total active power curve of the microgrid issued by the AI ​​server is decomposed into the active power adjustment range of the equipment to form a sequence of active power commands for the equipment, including:

[0080] Based on the feedback correction amount of the previous control cycle, the total active power curve of the microgrid within the rolling optimization window is target-corrected to obtain the total active power target for the current control period.

[0081] The rolling optimization window is the control time domain that starts from the current control cycle and covers subsequent control cycles. The strategy rolling decomposition module 102 extracts the target sequence within the rolling optimization window from the microgrid's total active power curve and applies the feedback correction amount from the previous control cycle to this target sequence to obtain the total active power target for the current control period. Target correction does not change the source and meaning of the microgrid's total active power curve issued by the AI ​​server 200: if there are no remaining tracking errors to be processed in the previous control cycle, the total active power target is determined by the target value corresponding to the microgrid's total active power curve; if there is a feedback correction amount, it is determined jointly by the target value and the feedback correction amount. Thus, the feedback correction amount simultaneously enters both the range calculation and target correction stages, which can reduce the accumulation of power target deviations between consecutive control cycles.

[0082] As one possible implementation method, the total active power target for the current control period is determined by the following formula:

[0083]

[0084] in, The total active power curve of the microgrid during the control period The corresponding target value, The feedback is generated by the closed-loop feedback module 104 and acts on the control cycle. The total amount of feedback correction is generated in the implementation of the closed-loop feedback module 104.

[0085] Within the range of the equipment's active power adjustment, the equipment's active power allocation quantity that satisfies the following formula is obtained. When multiple solutions exist, the equipment's active power command sequence is determined based on the energy conversion efficiency parameter and the power change constraints of adjacent control cycles:

[0086]

[0087] Where, in the formula The target total active power for the current control period. To control the cycle The set of edge devices that participate in the allocation. For edge devices The amount of active power allocation, To determine the candidate active power for the allocation of active power in the equipment, For edge devices The active power adjustment range of the equipment.

[0088] The strategy rolling decomposition module 102 selects candidate active power within the feasible region formed by the active power adjustment range of each participating edge device, minimizing the absolute value of the difference between the sum of the candidate active power and the total active power target. The resulting active power allocation meets the device-side executable conditions. If the adjustable capability of the participating edge devices can cover the total active power target, a sequence of active power commands matching the total active power target is formed. If it cannot fully cover the target, a sequence of active power commands with the smallest difference is formed, and the unfinished portion is handled by the closed-loop feedback process. When multiple solutions match the total active power target, the energy conversion efficiency parameter distinguishes the energy conversion differences when different edge devices undertake active power adjustment. The power change constraints of adjacent control cycles limit the command change amplitude of the same edge device within continuous control cycles. The solution that satisfies both requirements is prioritized, ensuring that the active power command sequence simultaneously meets the total target, the device executable range, and the continuous control stability requirements. Subsequently, the strategy rolling decomposition module 102 outputs this to the electrical constraint correction module 103.

[0089] As one possible implementation, when the candidate solution that minimizes the previous equation is not unique, the active power command sequence of the equipment is further determined within the set of candidate solutions using the secondary objective function shown in the following equation:

[0090]

[0091] in, The value of the secondary objective function. and These are preset weighting coefficients used to measure energy conversion losses and power changes in adjacent control cycles, respectively. These two coefficients are pre-configured based on operational requirements. For edge devices The energy conversion efficiency parameter, For edge devices The active power allocation in the previous control cycle.

[0092] In the electrical constraint correction module 103, the microgrid edge-side control terminal 100 reads the active power command sequence of the equipment and the node voltage, branch power, grid connection point power, and electrical operation limits provided by the grid-connected monitoring and control device 300. It should be noted that while the active power command sequence of the equipment numerically meets the total active power target, it does not necessarily meet the microgrid electrical operation constraints. Different edge devices are connected to different nodes, and the same active power adjustment has different effects on node voltage, branch power, and grid connection point power. Electrical constraint correction can reduce the risk of node voltage and branch power exceeding limits.

[0093] Specifically, the step of performing electrical constraint correction on the active power command sequence of the equipment based on real-time power flow calculation results and power sensitivity to form a corrected active power command sequence and power correction amount includes:

[0094] The active power command sequence of the device is mapped to the node injected power, and real-time power flow calculation is performed on the node injected power to obtain the real-time power flow calculation result;

[0095] The electrical constraint correction module 103 writes the active power allocation of each device in the active power command sequence into the corresponding node based on the microgrid network topology and edge device access node information. For nodes with multiple edge devices connected to the same node, the active power allocation of each device under that node is merged into the active power change of that node. For nodes not participating in the current control cycle allocation, the original injection state of the node is maintained based on the current operating data provided by the grid-connected monitoring and control device 300 and the data acquisition device 400, thus forming the node injection power. Subsequently, real-time power flow calculation is performed based on the node injection power, microgrid network topology, and line parameters to obtain the real-time power flow calculation results. The electrical quantities involved in the electrical operation limit verification include node voltage, branch power, and grid connection point power. This verification can detect node-side or line-side constraint conflicts that cannot be identified solely from the perspective of device capabilities.

[0096] Based on the real-time power flow calculation results and electrical operation limits, the electrical operation limit exceedance is determined, and the sensitivity of the correspondence between the node injected power and the limited electrical quantity is calculated to obtain the power sensitivity matrix.

[0097] The electrical constraint correction module 103 compares the real-time power flow calculation results with the electrical operating limits, which include the allowable range of node voltage, branch power, and grid connection point power. If the electrical quantity is within the corresponding electrical operating limit, there is no electrical operating violation; if the electrical quantity exceeds the corresponding electrical operating limit, the excess portion constitutes an electrical operating violation. Based on the network equations corresponding to the real-time power flow calculation results, the electrical constraint correction module 103 calculates the rate of change of the constrained electrical quantity relative to the injected power at each node, forming a power sensitivity matrix. The power sensitivity matrix represents the direction and magnitude of the influence of the active power adjustment of each edge device access node on the constrained electrical quantity, enabling subsequent corrections to determine the edge devices involved in the correction and their adjustment directions, reducing unnecessary changes in equipment commands.

[0098] As one possible implementation, the real-time power flow calculation is solved using the conventional Newton-Raphson method, or, in the case of a radial network structure, using the forward-backward substitution method. When the control period is short, a linearized power flow model can also be used. The power sensitivity matrix is ​​obtained by taking the partial derivatives of the microgrid node power equations at the current operating point. That is, the partial derivatives of node voltage, branch power, and grid connection point power relative to the active power injection of each node are extracted from the Jacobian matrix after the real-time power flow calculation converges, and the power sensitivity matrix is ​​formed according to the order of electrical quantities involved in the electrical operation limit verification. When the microgrid network topology or operating point changes, the power sensitivity matrix is ​​updated synchronously with the real-time power flow calculation results.

[0099] Among the power correction values ​​that make the corrected electrical quantities satisfy the following formula, the power correction value with the smallest power correction amplitude is selected, and the active power command sequence of the equipment is adjusted according to the power correction value to form the corrected active power command sequence of the equipment:

[0100]

[0101] Where, in the formula This refers to the electrical quantity vectors used in the electrical operation limit verification in the real-time power flow calculation results. and These are the lower and upper limit vectors of the electrical operating limits, respectively. This is the power sensitivity matrix corresponding to the current verification condition. This is the power correction amount. This is the active power vector corresponding to the active power command sequence of the equipment. This refers to the set of active power adjustment ranges for the equipment.

[0102] The power correction amount must simultaneously meet two conditions: first, the corrected electrical quantity is between the lower and upper limits of the electrical operating limits; second, the corrected active power vector of the equipment still belongs to the set of active power adjustment ranges of the equipment. Therefore, electrical constraint correction will not cause edge equipment to exceed its own active power adjustment range in order to meet node voltage or branch power constraints. The electrical constraint correction module 103 selects the power correction amount with the smallest power correction amplitude from the candidate power correction amounts that satisfy the above relationship. The power correction amplitude represents the degree of adjustment to the active power command sequence of the equipment. Selecting the power correction amount with the smallest power correction amplitude can retain the total active power decomposition result formed by the rolling decomposition of the strategy while meeting the electrical operating limits. Subsequently, the electrical constraint correction module 103 superimposes the power correction amount onto the active power command sequence of the equipment to form a corrected active power command sequence and records the power correction amount. Both are output to the closed-loop feedback module 104.

[0103] In the closed-loop feedback module 104, the microgrid edge-side control terminal 100 reads the corrected active power command sequence and power correction amount from the equipment, and reads the measured active power of the edge equipment transmitted back by the edge equipment controller 500, which is the actual power data formed after the edge equipment executes the corrected active power command sequence. It can be understood that the power correction amount is used to limit the attribution range of the equipment power tracking deviation. The adjustment made by the electrical constraint correction module 103 to the equipment active power command sequence is a planned adjustment and should not be treated as repeated compensation for deviations executed by the edge equipment. The resulting feedback correction amount reflects the true remaining tracking deviation and can reduce cross-cycle compensation errors.

[0104] like Figure 3 As shown, the closed-loop feedback module 104 includes: a deviation calculation unit 104-1, used to match the corrected equipment active power command sequence with the measured active power of the edge equipment according to the control time scale to obtain the equipment power tracking deviation;

[0105] In the deviation calculation unit 104-1, the microgrid edge-side control terminal 100 establishes a correspondence between the corrected command power and the measured active power of the same edge device within the same control period according to the device identifier and control time stamp, and calculates the difference between the two to obtain the device power tracking deviation. Matching according to the control time stamp can avoid deviation calculation errors caused by inconsistencies between the sampling time, communication return time, and command effective time of each edge device, ensuring that the device power tracking deviation accurately reflects the actual tracking of the edge device to the corrected device active power command sequence. The deviation calculation unit 104-1 outputs the device power tracking deviation to the feedback correction calculation unit 104-2.

[0106] Feedback correction calculation unit 104-2 is used to separate the planned deviation caused by the power correction from the power tracking deviation of the equipment, and to distribute the remaining tracking deviation across cycles according to the active power adjustment range of the equipment to obtain the feedback correction for the next control cycle.

[0107] In the feedback correction calculation unit 104-2, the microgrid edge control terminal 100 reads the power correction amount recorded by the electrical constraint correction module 103, subtracts the planned deviation caused by the power correction amount from the equipment power tracking deviation, and obtains the remaining tracking deviation. The planned deviation originates from the active power command sequence of the equipment made by the electrical constraint correction module 103 to meet the electrical operating limits. It is an in-plan adjustment and should not be used as repeated compensation for deviations performed by the edge equipment. Afterwards, the feedback correction calculation unit 104-2 performs cross-cycle allocation of the remaining tracking deviation according to the equipment active power adjustment range to obtain the feedback correction amount for the next control cycle, and outputs it to the write-back control unit 104-3.

[0108] The write-back control unit 104-3 is used to write the feedback correction amount for the next control cycle into the active power adjustment range calculation module 101 and the strategy rolling decomposition module 102.

[0109] In the write-back control unit 104-3, the microgrid edge-side control terminal 100 writes the feedback correction amount of the next control cycle into the equipment active power adjustment range calculation module 101 to correct the initial equipment active power adjustment range; at the same time, it writes it into the strategy rolling decomposition module 102 to correct the total active power target corresponding to the total active power curve of the microgrid within the rolling optimization window, so that the actual execution result of the previous control cycle enters the range calculation and target decomposition process of the next control cycle, forming a closed-loop feedback across control cycles.

[0110] Further, the power correction amount constrains the device power tracking deviation, generating a feedback correction amount for the next control cycle, including:

[0111] The corrected active power command sequence of the device is matched with the measured active power of the edge device according to the sampling period to obtain the device power tracking deviation;

[0112] Due to the time delay in device execution response, the measured active power of the edge device and the corrected active power command sequence need to be correlated according to the sampling period and control time scale. The closed-loop feedback module 104 matches the measured active power of the same edge device within the corresponding sampling period with the corresponding corrected command power based on the device identifier, control period, and sampling time scale to obtain the device power tracking deviation, thus avoiding errors in deviation calculation caused by inconsistent time scales.

[0113] After separating the planned deviation caused by the power correction amount from the power tracking deviation of the equipment, the remaining tracking deviation is allocated according to the active power adjustment range of the equipment in the subsequent control cycle to obtain the feedback correction amount for the next control cycle.

[0114] The planned deviation originates from the active adjustment of the equipment active power command sequence by the electrical constraint correction module 103. This adjustment is used to meet electrical operating limits and does not constitute a failure of the edge device execution. The closed-loop feedback module 104 subtracts the planned deviation from the equipment power tracking deviation based on the power correction amount to obtain the remaining tracking deviation. Then, the closed-loop feedback module 104 reads the equipment active power adjustment range in subsequent control cycles and allocates the remaining tracking deviation: if an edge device has a sufficient active power adjustment range to absorb the remaining tracking deviation in the next control cycle, the remaining tracking deviation is allocated to the next control cycle corresponding to that edge device; if it is insufficient to absorb all the remaining tracking deviation, the unabsorbed portion is allocated to other edge devices with adjustable capacity or to subsequent control cycles, forming the feedback correction amount for the next control cycle. This process avoids mistaking planned adjustments as execution deviation compensation and ensures that the actual execution deviation enters subsequent control cycles.

[0115] As one possible implementation method, edge devices are recorded. During the control cycle The active power allocation of the equipment is The power correction component is The corrected command power is the sum of the two. The measured active power is The deviation separation is then performed according to the following formula:

[0116]

[0117]

[0118] in, For the equipment power tracking deviation relative to the breakdown plan, That is, the planning deviation caused by the power correction amount. This represents the remaining tracking deviation after separation. The feedback correction for the next control cycle is generated using the following formula:

[0119]

[0120] And allocate across cycles according to the adjustable margin ratio of each edge device in the next control cycle:

[0121]

[0122] in, This represents the total amount of feedback correction for the next control cycle, with the unfinished power amount being positive. For edge devices During the control cycle The adjustable margin within the active power adjustment range of the equipment, consistent with the compensation direction. To allocate to edge devices The feedback correction component; when the adjustable margin is insufficient to absorb all the remaining tracking deviation, the unabsorbed portion is distributed to subsequent control cycles in the same manner.

[0123] In the local autonomous control module 105, the microgrid edge control terminal 100 continuously acquires the communication status with the AI ​​server 200: when the communication status is normal, the edge control is based on the total active power curve of the microgrid issued by the AI ​​server 200 and the corrected equipment active power command sequence formed after processing by the aforementioned modules; when the communication status is abnormal, the edge control is switched to the corrected equipment active power command sequence that has been most recently corrected by electrical constraints. This command sequence has been verified by real-time power flow calculation and electrical operation limits. Using it as the local active power reference can avoid stopping control due to the inability to receive new curves issued by the AI ​​server 200, thus improving the continuity of edge control.

[0124] Specifically, the step of switching edge-side control based on the communication status with the AI ​​server to obtain local control commands and synchronized operating status includes:

[0125] When the communication status with the AI ​​server is abnormal, the most recently corrected active power command sequence of the device through electrical constraint correction is used as the local active power reference, and the local active power reference is adjusted under the constraints of short-term source load power prediction and the active power adjustment range of the device to obtain local control commands.

[0126] The communication status is determined based on the data reception status, curve update time, and control command continuity with the AI ​​server 200. When the communication status is abnormal, the local autonomous control module 105 reads the local active power reference, short-term source-load power prediction, and equipment active power adjustment range. The short-term source-load power prediction is generated by the microgrid edge control terminal 100 based on historical photovoltaic output data and historical load power data provided by the data acquisition device 400, and is used to represent the short-term trend of photovoltaic output and load power changes. Under the above constraints, the local autonomous control module 105 adjusts the local active power reference to obtain local control commands, which are used to maintain the active power adjustment of edge devices during communication status abnormalities, ensuring continuous operation of edge-side control.

[0127] As one possible implementation method, during a communication status anomaly, the first... The total local active power target for each control cycle is determined by the following formula:

[0128]

[0129] in, This refers to the total active power corresponding to the local active power benchmark, which is the sum of the corrected equipment active power command sequence after the most recent electrical constraint correction. and These are the predicted changes in load power and photovoltaic output determined by short-term source load power forecasting, respectively. Subsequently, under the constraint of the active power adjustment range of each edge device, the total local active power target is decomposed to each edge device in the same allocation method as the strategy rolling decomposition module 102 to obtain local control commands.

[0130] After communication with the AI ​​server is restored, the local control commands are time-aligned with the operating measurements of the edge devices, and the total active power curve of the microgrid issued by the AI ​​server is synchronized to obtain a synchronized operating status.

[0131] After communication is restored, the local autonomous control module 105 first performs time alignment based on the execution timestamp of the local control commands and the sampling timestamp of the edge device's operating measurements to determine the actual active power status of the edge devices during each control cycle during the communication anomaly. Then, it synchronizes the total active power curve of the microgrid issued by the AI ​​server 200 and forms a synchronized operating state based on the edge device operating state at the time of communication restoration, the executed local control commands, and the total active power curve of the microgrid. The synchronized operating state is used for subsequent calculations of the device's active power adjustment range, strategy rolling decomposition, and electrical constraint correction, ensuring that the remote curve is aligned with the actual operating state on the edge side, avoiding sudden changes in edge device commands caused by directly switching control bases.

[0132] When the above modules are running, the equipment active power adjustment range calculation module 101 forms the equipment active power adjustment range, the strategy rolling decomposition module 102 forms the equipment active power command sequence within this range, the electrical constraint correction module 103 forms the corrected equipment active power command sequence and power correction amount, the closed-loop feedback module 104 generates the feedback correction amount for the next control cycle, and the local autonomous control module 105 switches the edge-side control basis according to the communication status and forms a synchronous operation state. Thus, the system forms a continuous control closed loop from equipment executable range calculation, strategy rolling decomposition, electrical constraint correction, closed-loop feedback to local autonomous control.

[0133] Example 2

[0134] like Figure 4As shown, based on the above system, this embodiment also provides a microgrid edge-side AI collaborative optimization control method. The parts not detailed in this embodiment are as shown in Embodiment 1. The method is applied to the microgrid edge-side control terminal 100 and includes:

[0135] S101: Calculate the active power adjustment range of the equipment based on the operating status of the edge device, the dynamic response model of the device, and the feedback correction amount of the previous control cycle.

[0136] S102: Decompose the total active power curve of the microgrid issued by the AI ​​server into the active power adjustment range of the device to form a sequence of active power instructions for the device;

[0137] S103: Perform electrical constraint correction on the active power command sequence of the equipment according to the real-time power flow calculation results and power sensitivity to form the corrected active power command sequence of the equipment and the power correction amount;

[0138] S104: Constrain the power tracking deviation of the equipment with the power correction amount, generate the feedback correction amount for the next control cycle, and use the feedback correction amount for the next control cycle as the input for the next calculation of the active power adjustment range of the equipment.

[0139] S105: When the communication status with the AI ​​server is abnormal, the most recently formed corrected device active power command sequence is used as the basis for local control, and after the communication status with the AI ​​server is restored, the status is synchronized to obtain local control commands and synchronized operation status.

[0140] The foregoing has only described certain exemplary embodiments of the present invention by way of illustration. Undoubtedly, those skilled in the art can modify the described embodiments in various ways without departing from the spirit and scope of the present invention. Therefore, the foregoing drawings and descriptions are illustrative in nature and should not be construed as limiting the scope of protection of the claims of the present invention.

Claims

1. A microgrid edge-side AI collaborative optimization control system, characterized in that, include: The active power adjustment range calculation module is used to calculate the active power adjustment range of the equipment based on the operating status of the edge equipment, the dynamic response model of the equipment, and the feedback correction amount of the previous control cycle. The strategy rolling decomposition module is used to decompose the total active power curve of the microgrid issued by the AI ​​server into the active power adjustment range of the device, forming a sequence of active power command for the device. The electrical constraint correction module is used to perform electrical constraint correction on the active power command sequence of the equipment according to the real-time power flow calculation results and power sensitivity, so as to form the corrected active power command sequence of the equipment and the power correction amount; The closed-loop feedback module is used to constrain the power tracking deviation of the equipment with the power correction amount, generate the feedback correction amount for the next control cycle, and use the feedback correction amount for the next control cycle as the input of the active power adjustment range calculation module of the equipment. The local autonomous control module is used to switch the edge-side control basis according to the communication status with the AI ​​server, and to obtain local control commands and synchronous operation status.

2. The system according to claim 1, characterized in that, The device dynamic response model includes active power response relationship, operating boundary constraints and energy conversion efficiency parameters; the active power response relationship is characterized by active power ramp-up limit and response delay, the operating boundary constraints are characterized by rated active power range and state quantity allowable range, and the energy conversion efficiency parameters are used to constrain the active power allocation of the device when the total active power curve of the microgrid is decomposed to the edge device.

3. The system according to claim 1, characterized in that, The calculation of the active power adjustment range of the equipment based on the operating status of the edge device, the device dynamic response model, and the feedback correction amount of the previous control cycle includes: The operational measurement values ​​of edge devices are time-aligned according to the control cycle to obtain the cycle operation status; The initial active power adjustment range of the equipment is determined based on the cycle operation status and the equipment dynamic response model; The feedback correction from the previous control cycle is applied to the initial active power adjustment range of the equipment, and the correction result is restricted within the operating boundary constraints to obtain the active power adjustment range of the equipment.

4. The system according to claim 3, characterized in that, The process of determining the initial active power adjustment range of the equipment based on the periodic operating state and the equipment dynamic response model includes: Based on the equipment state variables and operating boundary constraints in the cycle operation state, the upper and lower limits of the active power that can be output or absorbed in the current control cycle are determined, and the state-allowed power range is obtained. Based on the current measured active power and active power response relationship in the cycle operation state, the upper and lower limits of active power that can be reached in the next control cycle are determined, and the ramp constraint power range is obtained. After mapping the state-allowed power range and the ramp-constrained power range to the same control time period, the intersection is calculated, and the initial equipment active power adjustment range is obtained by the following formula: Where, in the formula For edge devices During the control cycle The initial active power adjustment range of the equipment. and These are the lower and upper limits of the allowable power range for the stated state, respectively. and These are the lower and upper limits of the climbing constraint power range, respectively.

5. The system according to claim 1, characterized in that, The step of decomposing the total active power curve of the microgrid issued by the AI ​​server into the active power adjustment range of the equipment to form a sequence of active power commands for the equipment includes: Based on the feedback correction amount of the previous control cycle, the total active power curve of the microgrid within the rolling optimization window is target-corrected to obtain the total active power target for the current control period. Within the range of the equipment's active power adjustment, the equipment's active power allocation quantity that satisfies the following formula is obtained. When multiple solutions exist, the equipment's active power command sequence is determined based on the energy conversion efficiency parameter and the power change constraints of adjacent control cycles: Where, in the formula The target total active power for the current control period. To control the cycle The set of edge devices that participate in the allocation. For edge devices The amount of active power allocation, To determine the candidate active power for the allocation of active power in the equipment, For edge devices The active power adjustment range of the equipment.

6. The system according to claim 1, characterized in that, The process of performing electrical constraint correction on the active power command sequence of the equipment based on real-time power flow calculation results and power sensitivity, to form a corrected active power command sequence and power correction amount, includes: The active power command sequence of the device is mapped to the node injected power, and real-time power flow calculation is performed on the node injected power to obtain the real-time power flow calculation result; Based on the real-time power flow calculation results and electrical operation limits, the electrical operation limit exceedance is determined, and the sensitivity of the correspondence between the node injected power and the limited electrical quantity is calculated to obtain the power sensitivity matrix. Among the power correction values ​​that make the corrected electrical quantities satisfy the following formula, the power correction value with the smallest power correction amplitude is selected, and the active power command sequence of the equipment is adjusted according to the power correction value to form the corrected active power command sequence of the equipment: Where, in the formula This refers to the electrical quantity vectors used in the electrical operation limit verification in the real-time power flow calculation results. and These are the lower and upper limit vectors of the electrical operating limits, respectively. This is the power sensitivity matrix corresponding to the current verification condition. This is the power correction amount. This is the active power vector corresponding to the active power command sequence of the equipment. This refers to the set of active power adjustment ranges for the equipment.

7. The system according to claim 1, characterized in that, The closed-loop feedback module includes: a deviation calculation unit, used to match the corrected equipment active power command sequence with the measured active power of the edge equipment according to the control time scale to obtain the equipment power tracking deviation; a feedback correction calculation unit, used to separate the planned deviation caused by the power correction from the equipment power tracking deviation, and to allocate the remaining tracking deviation across cycles according to the equipment active power adjustment range to obtain the feedback correction for the next control cycle; and a write-back control unit, used to write the feedback correction for the next control cycle into the equipment active power adjustment range calculation module and the strategy rolling decomposition module.

8. The system according to claim 1, characterized in that, The step of constraining the power tracking deviation of the device with the power correction amount and generating the feedback correction amount for the next control cycle includes: The corrected active power command sequence of the device is matched with the measured active power of the edge device according to the sampling period to obtain the device power tracking deviation; After separating the planned deviation caused by the power correction amount from the power tracking deviation of the equipment, the remaining tracking deviation is allocated according to the active power adjustment range of the equipment in the subsequent control cycle to obtain the feedback correction amount for the next control cycle.

9. The system according to claim 1, characterized in that, The method of switching edge-side control based on the communication status with the AI ​​server to obtain local control commands and synchronized operating status includes: When the communication status with the AI ​​server is abnormal, the most recently corrected active power command sequence of the device through electrical constraint correction is used as the local active power reference, and the local active power reference is adjusted under the constraints of short-term source load power prediction and the active power adjustment range of the device to obtain local control commands. After communication with the AI ​​server is restored, the local control commands are time-aligned with the operating measurements of the edge devices, and the total active power curve of the microgrid issued by the AI ​​server is synchronized to obtain a synchronized operating status.

10. A microgrid edge-side AI collaborative optimization control method, characterized in that, include: S101: Calculate the active power adjustment range of the equipment based on the operating status of the edge device, the dynamic response model of the device, and the feedback correction amount of the previous control cycle. S102: Decompose the total active power curve of the microgrid issued by the AI ​​server into the active power adjustment range of the device to form a sequence of active power instructions for the device; S103: Perform electrical constraint correction on the active power command sequence of the equipment according to the real-time power flow calculation results and power sensitivity to form the corrected active power command sequence of the equipment and the power correction amount; S104: Constrain the power tracking deviation of the equipment with the power correction amount, generate the feedback correction amount for the next control cycle, and use the feedback correction amount for the next control cycle as the input for the next calculation of the active power adjustment range of the equipment. S105: When the communication status with the AI ​​server is abnormal, the most recently formed corrected device active power command sequence is used as the basis for local control, and after the communication status with the AI ​​server is restored, the status is synchronized to obtain local control commands and synchronized operation status.

Images