Distributed power supply power coordinated optimization method and system based on edge calculation

Abstract

The invention discloses a distributed power supply power coordinated optimization method and system based on edge calculation, and relates to the technical field of intelligent power grid optimization control, and the method comprises the steps: collecting operation parameters at an edge node, building a short-period power prediction model, calculating the power generation capability and load evolution trend, and combining the equipment working condition, load characteristics and historical operation rules, outputting a power output plan and a pre-distribution value; a lightweight communication mechanism is introduced, a short-period power prediction model is shared among edge nodes to predict an output value, an operation load and residual regulation capability, and local correction is performed on output power and load pressure; and after local correction, collecting current power grid fluctuation and economic signal data, outputting a final coordinated and optimized scheduling scheme, continuously monitoring an execution power output effect and a prediction deviation, and carrying out emergency compensation by adjusting the scheduling scheme. According to the method, real-time prediction, collaborative optimization and dynamic linkage of the power grid of the distributed power supply are realized, and the adaptability and energy efficiency optimization of the power grid are enhanced.

Description

Technical Field

[0001] The present invention relates to the field of smart grid optimization control technology, and specifically to a distributed power supply power coordination optimization method and system based on edge computing. Background Art

[0002] With the large-scale integration of renewable energy into power systems, the proportion of distributed energy resources (DER) in modern power grids continues to increase, forming a highly diverse and dynamically changing power supply landscape. Traditional centralized dispatch and control architectures suffer from long response delays and coarse dispatch granularity when faced with the large-scale integration of DERs. In recent years, edge computing, a new computing architecture with local real-time processing capabilities, has been gradually applied to smart grids, enabling autonomous control of DERs. Real-time data collection of operating status by edge nodes, the construction of short-term prediction models, and collaborative optimization of local power dispatch have become important technical directions for promoting the efficient operation of new power systems. Furthermore, the introduction of lightweight communication mechanisms and local collaborative algorithms is exploring alternatives to traditional centralized control methods, enhancing system flexibility and resilience.

[0003] While some research and applied practices have explored distributed power generation optimization and control methods based on edge computing, several key deficiencies remain. First, existing technologies mostly focus on static feature modeling and lack short-term prediction-driven mechanisms for rapidly changing scenarios (such as fluctuations in renewable energy output and drastic load changes). This makes it impossible to generate high-frequency, high-precision power output plans, resulting in scheduling plans lagging behind the actual system state. Second, currently, most approaches utilize centralized coordination or global model-based optimization algorithms, which incur high communication overhead and lack lightweight, real-time collaborative optimization capabilities for edge nodes. This can easily lead to communication bottlenecks and computational burdens in large-scale, multi-node scenarios. Furthermore, existing solutions generally neglect the design of dynamic interaction with the main grid. They lack adaptive coordination and control mechanisms driven by real-time grid fluctuation data (such as frequency, electricity prices, and voltage), making it impossible to ensure dynamic consistency and balance between the system and the main grid. Summary of the Invention

[0004] In view of the above-mentioned problems, the present invention is proposed.

[0005] Therefore, the technical problem solved by the present invention is: the existing distributed power supply coordination optimization method lacks a high-frequency short-cycle power prediction driving mechanism, lacks lightweight real-time collaborative optimization capabilities, fails to realize the dynamic linkage control mechanism of the power grid, and how to construct an end-to-end real-time adaptive distributed power supply power coordination optimization method.

[0006] To solve the above technical problems, the present invention provides the following technical solutions: a distributed power supply coordination and optimization method based on edge computing, comprising collecting operating parameters at edge nodes, constructing a short-cycle power prediction model, calculating the power generation capacity and load evolution trend, and outputting a power output plan and pre-allocated value in combination with equipment operating conditions, load characteristics and historical operating rules; introducing a lightweight communication mechanism to share the predicted output value, operating load and remaining adjustment capacity of the short-cycle power prediction model between edge nodes, and performing local correction on the output power and load pressure; collecting the current power grid fluctuation and economic signal data after local correction, outputting the final coordinated and optimized scheduling plan, and continuously monitoring the execution power output effect and prediction deviation, and performing emergency compensation by adjusting the scheduling plan.

[0007] As an optimal solution of the distributed power supply power coordination optimization method based on edge computing described in the present invention, the operating parameters include equipment operating status data, historical power and trend data, load status and evolution data, energy storage device status data, external environment disturbance data and main grid interface preliminary synchronization data.

[0008] As an optimal solution of the distributed power supply power coordination optimization method based on edge computing described in the present invention, the construction of a short-cycle power prediction model includes using historical power sequences, environmental disturbance factors and load evolution trends as input variables to construct a short-cycle power prediction model, and output the prediction of power generation capacity and load trends in the future scheduling cycle.

[0009] As a preferred solution of the distributed power supply power coordination optimization method based on edge computing described in the present invention, the lightweight communication mechanism includes sharing and transmitting short-period predicted output values, operating load information and remaining regulation capabilities between edge nodes based on adjacent node topology through a point-to-point low-overhead data exchange protocol.

[0010] As a preferred solution of the distributed power supply power coordination optimization method based on edge computing described in the present invention, the local correction of output power and load pressure includes introducing the load difference, power difference and adjustment margin between the current node and the adjacent node, and dynamically adjusting the power output plan of each node based on the local collaborative optimization algorithm.

[0011] As a preferred solution of the distributed power supply power coordination optimization method based on edge computing described in the present invention, the current power grid fluctuation and economic signal data include power grid frequency, power grid interface voltage, real-time electricity price, and power purchase and feed power information.

[0012] As a preferred solution of the distributed power supply power coordination optimization method based on edge computing described in the present invention, the continuous monitoring of the execution power output effect and the predicted deviation includes real-time collection of the actual power output of each node, the deviation rate between the output and the power output plan, and triggering emergency compensation measures according to the deviation level.

[0013] Another object of the present invention is to provide a distributed power supply coordination and optimization system based on edge computing, which can collect operating parameters at the edge node, build a short-cycle power prediction model, calculate the power generation capacity and load evolution trend, and output power output plan and pre-allocated value in combination with equipment operating conditions, load characteristics and historical operating rules, thereby solving the problem that the current distributed power supply coordination optimization method lacks a high-frequency short-cycle power prediction driving mechanism.

[0014] As an optimal solution of the distributed power supply power coordination optimization system based on edge computing described in the present invention, it includes: an edge node state perception and short-cycle power prediction module, an edge node collaborative optimization and local correction module, and a power grid interactive regulation and closed-loop optimization feedback module; the edge node state perception and short-cycle power prediction module is used to collect operating parameters at each edge node in real time, build a short-cycle power prediction model based on the collected data, calculate future power generation capacity and load evolution trend, and output local power output plan and preliminary pre-allocated value in combination with historical operating rules and load characteristics; the edge node collaborative optimization and local correction module is used to share the respective short-cycle predicted output values, operating loads and remaining regulation capabilities between edge nodes through a lightweight communication mechanism, and use the shared data to perform local corrections to the power output plan of each node; the power grid interactive regulation and closed-loop optimization feedback module is used to perform final coordinated optimization adjustment on the power output of each node, and continuously monitor the power output effect and prediction deviation during the scheduling execution stage.

[0015] A computer device includes a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement a step of a distributed power supply power coordination optimization method based on edge computing.

[0016] A computer-readable storage medium stores a computer program, which, when executed by a processor, implements the steps of a distributed power supply power coordination optimization method based on edge computing.

[0017] Beneficial effects of the present invention: The distributed power supply power coordination optimization method based on edge computing provided by the present invention collects a wealth of operating parameters at the edge nodes, constructs a short-cycle power prediction model, calculates the power generation capacity and load evolution trend of the distributed power supply in real time, and combines the equipment working conditions, load characteristics and historical operating rules to achieve forward-looking output power planning and reasonable pre-allocation values, providing high-quality data support for the dynamic scheduling of the entire system, thereby effectively reducing scheduling lags and operational risks. Furthermore, by introducing a lightweight communication mechanism, efficient sharing of short-cycle predicted output values, operating loads and remaining regulation capabilities can be achieved between edge nodes, local state information can be dynamically synchronized, and the power output of each node can be adjusted through a local correction mechanism, thereby optimizing the power distribution within the region, reducing power conflicts and load imbalances between nodes, and enhancing the distributed autonomy and flexibility of the system. At the same time, on the basis of local correction, the dynamic fluctuations and economic signals of the frequency, voltage, electricity price and purchased / fed power of the main power grid are collected in real time, and a coordinated and optimized dispatch plan is further output. By continuously monitoring the power execution effect and prediction deviation, dynamic adjustment and emergency compensation are implemented. This can ensure the dynamic consistency of the overall system with the main power grid, improve the adaptability to grid disturbances, enhance the stability and economy of power interaction, and finally construct a distributed power supply power coordination optimization method with real-time prediction drive, local collaborative optimization and global dynamic closed-loop control capabilities, which is significantly better than the existing centralized control or static optimization mode. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0019] Figure 1 An overall flow chart of a distributed power supply power coordination optimization method based on edge computing provided in the first embodiment of the present invention.

[0020] Figure 2 A logic flow chart of a distributed power supply power coordination optimization method based on edge computing provided in the first embodiment of the present invention.

[0021] Figure 3 An overall flow chart of a distributed power supply power coordination and optimization system based on edge computing provided in the third embodiment of the present invention. DETAILED DESCRIPTION

[0022] To make the above-mentioned objects, features, and advantages of the present invention more clearly understood, the following detailed description of the specific embodiments of the present invention is given in conjunction with the accompanying drawings. It is obvious that the described embodiments are only part of the embodiments of the present invention, not all of them. Based on the embodiments of the present invention, all other embodiments obtained by ordinary persons in this field without creative work should fall within the scope of protection of the present invention.

[0023] Example 1, with reference to Figure 1-Figure 2 , as an embodiment of the present invention, provides a distributed power supply power coordination optimization method based on edge computing, including: S1: Collect operating parameters at the edge node, build a short-cycle power prediction model, calculate the power generation capacity and load evolution trend, and output the power output plan and pre-allocated value based on the equipment operating conditions, load characteristics and historical operating rules.

[0024] Furthermore, for each distributed power supply node, operating parameters are sensed and collected in real time, including: equipment operating status data, historical power and trend data, load status and evolution data, energy storage device status data, external environmental disturbance data, and preliminary synchronization data of the main grid interface.

[0025] Historical power and trend data includes current instantaneous output power, device voltage, device current, device frequency, device operating mode status, and device operating temperature.

[0026] Historical power and trend data include power series within the historical time window period, sliding average power and historical prediction error records.

[0027] Load status and evolution data include current node load, load classification data (adjustable load / rigid load / delayable load), real-time load change rate, and predicted future load trends (which can be collected if a load forecasting model is available).

[0028] Energy storage device status data includes the current energy storage charge state, energy storage charge and discharge power, current energy storage mode (charging / discharging / standby) and charge change rate.

[0029] External environmental disturbance data include: for wind power nodes: current wind speed, wind direction and wind speed change rate; for photovoltaic nodes: current light intensity, panel temperature and light change rate.

[0030] The preliminary synchronization data of the main grid interface includes the grid interface voltage, grid interface frequency, grid electricity price, and current purchased / fed power.

[0031] It should be noted that based on the collected operating parameters, the edge device performs short-term power forecasting, calculates the power generation capacity and load evolution trend in the future, and constructs a short-term power forecasting model expressed as:

[0032] in, Predict the power value for the next cycle, is the historical power mean deviation per unit time, is the square of the forecast error of the previous period, is the load change rate of the current cycle, is the environmental disturbance factor (such as wind speed change rate), is the disturbance response adjustment coefficient, is the history matching stability parameter, To predict the master amplification factor, is the load-adjusted damping coefficient, is the current load curve function, Indicates time.

[0033] like The range is , indicating that the node is predicted to be in low output or standby shutdown; if The range is , indicating that the node is in the normal load operation range; if The range is , indicating that the node is in a high load / maximum resource utilization state.

[0034] Each node generates a local preliminary power output plan based on the equipment operating conditions, load characteristics, and historical operating rules, forming a pre-allocated value expressed as:

[0035] in, For the Nodes in The power output plan value of the cycle, is the short-term power generation power predicted for the node, The current equipment operating capacity factor represents the ratio of the equipment's allowable output capacity to its rated capacity. is the load disturbance sensitivity coefficient, which indicates the device's ability to suppress severe load fluctuations. It is the rate of change of the node load in the current cycle, which is used to dynamically reflect the load increase or decrease trend.

[0036] when (load stability), the denominator approaches 1, and the planned value approaches the equipment capacity ratio of the predicted value, indicating that the output is released according to the predicted results; when Increase (severe load fluctuations), the denominator increases, and the system automatically reduces power output to avoid rapid and frequent adjustments of equipment; , the actual equipment output capacity is updated in real time according to the load and status; the output power is always , which has the function of regulating margin control.

[0037] S2: A lightweight communication mechanism is introduced to share the short-period power prediction model among edge nodes to predict output values, operating loads, and remaining regulation capabilities, and to perform local corrections on output power and load pressure.

[0038] Furthermore, after completing the local short-cycle power forecast and output plan formulation, in order to prevent power conflicts or load offset problems caused by the isolated operation of each distributed power supply node, the system builds a lightweight asynchronous communication mechanism between edge nodes to realize the shared transmission of power plans, load status and regulation capabilities.

[0039] Using a Local Adjacency Graph (LAG) structure, each node only needs to exchange state data with its physical or scheduling neighbors, significantly reducing communication bandwidth pressure and synchronization difficulties. Within a fixed scheduling cycle, each edge node encapsulates and sends its own state (including predicted power value, current load, remaining available power, load change rate, etc.) to adjacent nodes via a broadcast message.

[0040] After receiving data from neighboring nodes, each node makes a local correction based on the shared information and local plan: if its own power output is significantly higher than the average predicted power of the neighborhood, the output is reduced; if it is in an environment with high load pressure and strong neighborhood redundancy, the planned power is appropriately increased; if drastic load fluctuations at multiple points are detected, energy storage support is called.

[0041] All nodes in the system achieve rapid convergence to a power balance state through local adjustment mechanisms based on negotiation criteria, such as linear correction method, collaborative gradient descent method or simple distributed ADMM, avoiding dependence on the central controller and significantly improving the system fault tolerance and adjustment robustness.

[0042] Furthermore, the following main grid parameters are first acquired in real time through the grid interface acquisition module: the main grid frequency is used to detect frequency drift and overfrequency / underfrequency behavior; the main grid electricity price is used to achieve economic optimization of power response; the interface point voltage is used to determine whether there is a voltage drop or backfeed risk; the system purchased / fed power measures the size of the power interaction between the system and the main grid.

[0043] The data collection frequency is adjustable, and the preferred value is once every 1 to 5 seconds.

[0044] It should be noted that after receiving the main grid status information, it is jointly analyzed with the generated local power output plan and node redundancy capabilities to form the following linkage response strategy: If the main grid frequency is greater than the normal frequency , (overfrequency), then reduce the output of non-critical load nodes; if the main grid frequency is lower than the normal frequency (underfrequency), nodes with energy storage capabilities are mobilized to quickly increase power to support the main network.

[0045] When the main grid electricity price Above the set threshold , and when the system has sufficient redundancy, the system automatically reduces electricity purchases and increases the proportion of independent energy supply; - When the electricity price is too low (such as during the period of clean electricity surplus), the system automatically absorbs external power to charge local energy storage.

[0046] If the interface voltage When the set critical value is approached, the system adjusts the output plan to make the power tend to be balanced and avoid protection actions caused by voltage drop or reverse flow.

[0047] Before executing the planned power output, the system introduces the main grid response factor and makes a secondary adjustment to the power allocation value of each node, which is expressed as:

[0048] in, The main network response sensitivity coefficient is set by the system; : is the "participation index" of node i in the grid response in the current scheduling cycle, which is obtained by the system by weighting factors such as its redundancy capability, response capability and location voltage value; if the node has energy storage, the system can automatically calculate the "standby release capacity" according to the energy storage state of charge (SOC) to participate in grid regulation.

[0049] S3: After local correction, the current power grid fluctuation and economic signal data are collected, and the final coordinated and optimized dispatch plan is output. The executed power output effect and the forecast deviation are continuously monitored, and emergency compensation is performed by adjusting the dispatch plan.

[0050] Furthermore, the system first sends the final node power output value to each edge node controller, and the device side controls the actual power output process according to the accuracy level and adjustment capability.

[0051] During the output process, the system continuously collects the following key execution data: actual output power; node response delay; real-time frequency offset; actual energy storage discharge / charging curve; and load satisfaction rate of this round of execution (used to determine whether the scheduling accurately meets the demand).

[0052] It should be noted that the system quantifies the deviation between the actual power and the planned value and calculates the relative deviation rate:

[0053] in, For the Edge nodes in Relative deviation rate of power output during the scheduling period; For the Edge nodes in The actual power output value during the scheduling period; For the Edge nodes in The target power output value calculated according to the final coordination optimization plan during the dispatch period; A small positive constant introduced to prevent the denominator from being zero.

[0054] when , which is a Level I (mild) log and does not trigger correction; when , which is level II (moderate), start local plan re-smoothing; when Level III (serious) triggers neighborhood renegotiation and uploads cloud scheduling reconstruction.

[0055] Furthermore, if a moderate or severe deviation level is triggered, the system will implement the following two-level compensation strategy: Local rapid correction compensation: If the energy storage within the node still has discharge capacity or some loads can be started later, the system will perform local compensation, using energy storage to peak load or delaying load operation to bridge the gap; Neighborhood coordinated resource compensation: If local adjustment is not possible, power support is requested through neighboring node status broadcasts, and local compensation adjustments are made within the scheduling to avoid system-wide disturbances.

[0056] It should be noted that after each complete scheduling cycle, the system will upload the following structured data packets to the central cloud: predicted value, execution value, deviation rate; scheduling plan and final actual behavior comparison matrix; energy storage behavior curve and system power fluctuation event record; node compensation behavior trajectory (including source and response time); the cloud platform will perform the following optimization behaviors based on this: optimize the prediction model parameters; re-evaluate the node level and adjust the priority weight; rebuild the next round of scheduling weight matrix after detecting long-term system behavior deviations; and form templates for some "excellent response" behaviors and distribute them to the entire system for promotion.

[0057] When the system has an average deviation rate of 3 cycles , or a key node reaches level III deviation twice in a row, the whole system scheduling reconstruction mechanism is triggered.

[0058] Example 2, an embodiment of the present invention, provides a distributed power supply power coordination optimization method based on edge computing. In order to verify the beneficial effects of the present invention, scientific demonstration is carried out through economic benefit calculation and simulation experiments.

[0059] First, a microgrid in an industrial park was selected as the experimental subject. This microgrid consists of five distributed power generation nodes: three photovoltaic nodes (PV1-PV3), one wind power node (WT1), and one energy storage node (ESS1). All nodes are connected to the main grid and have real-time grid-connected capabilities. Edge computing devices are deployed on each node, providing local data collection and predictive computing capabilities, as well as enabling neighborhood communication and interconnection. The experiment lasted five consecutive days, with six 10-minute scheduling cycle data collected daily.

[0060] Each edge node collects operating parameters, including device operating status (real-time power, voltage, current, frequency, device status, and temperature), historical power series, sliding average power, load classification and rate of change, energy storage charge state and rate of change, environmental disturbances (light intensity, wind speed), and main grid interface voltage, frequency, electricity price, and purchased / fed-in power. The data is collected every one minute and fed into a short-term power forecasting model to predict power trends for the next 20 minutes and generate node-level power output plans and pre-allocated values. To ensure forecast accuracy, the model dynamically adjusts the forecast window and parameters. Edge nodes share short-term forecast results, load status, and remaining regulation capacity via a lightweight communication protocol (based on MQTT), construct a local adjacency graph, and employ a collaborative correction strategy to adjust each node's power output plan. Nodes communicate only with physically adjacent nodes to avoid communication congestion. Local correction results are updated during each scheduling cycle. The system collects main grid fluctuation signals in real time and dynamically adjusts the final power output plan based on electricity price, frequency, and voltage. During execution, the system monitors actual power output, response delay, energy storage charging and discharging behavior, and load satisfaction in real time, calculates power deviation rates, and implements compensation strategies in a tiered manner. If the deviation is too large, neighboring nodes automatically coordinate compensation, and system data is synchronously uploaded to the cloud to optimize the prediction model and scheduling weights.

[0061] Table 1 Experimental data table

[0062] The experimental data in Table 1 demonstrates that, through the implementation of the present method, each node in the distributed power system demonstrates significant coordination and dynamic adaptability during short-term power forecasting and collaborative optimization. First, the adjustment between the predicted power and the revised planned power demonstrates the effectiveness of neighborhood collaborative correction achieved through lightweight communication in the S2 phase. For example, the original predicted power of node WT1 was 30 kW, which was increased to 32 kW after revision, and the actual output reached 33 kW. This indicates that, upon detecting strong neighborhood redundancy, the system proactively increased the power output of this node, enhancing local load balancing capabilities. Energy storage node ESS1 intelligently switched to a discharge state based on the main grid frequency and electricity price signals. The revised discharge power plan was -10 kW, and the actual discharge power was 9.5 kW. This demonstrates that the present method possesses high dynamic energy storage control capabilities and can effectively support optimization strategies during periods of main grid underfrequency or high electricity prices.

[0063] In terms of deviation rate, the deviation rate of all nodes is controlled within 5%, and most of the nodes remain in the range of 2% to 3%, indicating that the present invention has significantly improved the accuracy of power plan execution through the real-time monitoring and emergency compensation mechanism of the S3 stage. Compared with the traditional centralized scheduling method, the latter usually relies on static models and periodic global plan issuance, lacks short-cycle correction and edge collaboration, and easily leads to power deviations as high as 8% to 12%, especially in high-volatility new energy scenarios. This embodiment uses a closed-loop control system of edge prediction + collaborative optimization + grid dynamic linkage + feedback compensation to achieve stable power output under dynamic conditions, reduce the system's dependence on the central controller, and improve scheduling flexibility and system resilience.

[0064] Furthermore, combining the main grid electricity price (0.75 yuan / kWh) with the frequency signal (maintained in the average range of 49.95-50.02 Hz), it can be seen that the system has good frequency support and economic optimization capabilities. In high electricity price scenarios, it actively reduces the proportion of electricity purchased from the main grid, enhances local absorption capacity, and optimizes grid frequency support through energy storage, further improving the quality of power interaction. In summary, the examples fully verify the innovation and practical application value of the method of the present invention in dynamic power forecasting, distributed collaborative optimization, real-time grid interaction, and closed-loop feedback control, significantly outperforming existing centralized control and static scheduling technology systems.

[0065] Example 3, reference Figure 3 This paper, an embodiment of the present invention, provides a distributed power coordination and optimization system based on edge computing, including an edge node state perception and short-term power prediction module, an edge node collaborative optimization and local correction module, and a grid interactive control and closed-loop optimization feedback module. Among them, the edge node state perception and short-cycle power prediction module is used to collect operating parameters at each edge node in real time, build a short-cycle power prediction model based on the collected data, calculate future power generation capacity and load evolution trend, and combine historical operating rules and load characteristics to output local power output plan and preliminary pre-allocation value. The edge node collaborative optimization and local correction module is used to share the respective short-cycle predicted output values, operating loads and remaining adjustment capabilities between edge nodes through a lightweight communication mechanism, and use shared data to make local corrections to the power output plan of each node. The grid interactive control and closed-loop optimization feedback module is used to make final coordinated optimization adjustments to the power output of each node, and continuously monitor the power output effect and prediction deviation during the scheduling execution phase.

[0066] If a function is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, or the portion that contributes to the prior art, or the portion 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 for causing a computer device (which can be a personal computer, server, or network device, etc.) to execute all or part of the steps of the various embodiments of the method of the present invention. The aforementioned storage medium includes various media that can store program code, such as USB flash drives, mobile hard drives, read-only memories (ROMs), random access memories (RAMs), magnetic disks, or optical disks.

[0067] The logic and / or steps represented in the flowcharts or otherwise described herein, for example, can be considered as a sequenced list of executable instructions for implementing the logical functions, and can be embodied in any computer-readable medium for use by, or in conjunction with, an instruction execution system, apparatus, or device (e.g., a computer-based system, a system including a processor, or other system that can fetch and execute instructions from an instruction execution system, apparatus, or device). For purposes of this specification, a "computer-readable medium" can be any device that can contain, store, communicate, propagate, or transport a program for use by, or in conjunction with, an instruction execution system, apparatus, or device.

[0068] More specific examples (a non-exhaustive list) of computer-readable media include the following: an electrical connection with one or more wires (electronic devices), a portable computer disk cartridge (magnetic devices), a random access memory (RAM), a read-only memory (ROM), an erasable and programmable read-only memory (EPROM or flash memory), a fiber optic device, and a portable compact disc read-only memory (CDROM). In addition, the computer-readable medium may even be paper or other suitable medium on which the program is printed, since the program may be obtained electronically, for example, by optically scanning the paper or other medium and then editing, interpreting, or processing it in another suitable manner as necessary, and then storing it in a computer memory.

[0069] It should be understood that various aspects of the present invention may be implemented using hardware, software, firmware, or a combination thereof. In the above-described embodiments, multiple steps or methods may be implemented using software or firmware stored in a memory and executed by a suitable instruction execution system. For example, if implemented using hardware, as in another embodiment, any one or a combination of the following technologies known in the art may be used: a discrete logic circuit having logic gates for implementing logic functions on data signals, an application-specific integrated circuit having suitable combinational logic gates, a programmable gate array (PGA), a field-programmable gate array (FPGA), etc. It should be noted that the above embodiments are intended only to illustrate the technical solutions of the present invention and are not intended to be limiting. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art will understand that modifications or equivalent substitutions may be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention, and such modifications are intended to be encompassed by the claims of the present invention.

[0070] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit the present invention. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical solutions of the present invention, which should all be included in the scope of the claims of the present invention.

Claims

1. A distributed power supply power coordination optimization method based on edge computing, characterized in that: include: Collect operating parameters at the edge node, build a short-term power prediction model, calculate power generation capacity and load evolution trends, and output power output plans and pre-allocated values ​​based on equipment operating conditions, load characteristics, and historical operating patterns; A lightweight communication mechanism is introduced to share the output value, operating load, and remaining regulation capacity of the short-term power prediction model between edge nodes, and to perform local corrections on output power and load pressure. After local correction, the current power grid fluctuation and economic signal data are collected, and the final coordinated and optimized dispatch plan is output. The execution power output effect and forecast deviation are continuously monitored, and emergency compensation is carried out by adjusting the dispatch plan.

2. The distributed power supply power coordination optimization method based on edge computing according to claim 1, characterized in that: The operating parameters include equipment operating status data, historical power and trend data, load status and evolution data, energy storage equipment status data, external environmental disturbance data, and main grid interface preliminary synchronization data.

3. The distributed power supply power coordination optimization method based on edge computing according to claim 2, characterized in that: The short-term power prediction model is constructed by using historical power series, environmental disturbance factors and load evolution trends as input variables to construct a short-term power prediction model and output a prediction of power generation capacity and load trends in the future scheduling cycle.

4. The distributed power supply power coordination optimization method based on edge computing according to claim 3 is characterized in that: The lightweight communication mechanism includes sharing and transmitting short-period predicted output values, operating load information and remaining regulation capabilities between edge nodes based on adjacent node topology through a point-to-point low-overhead data exchange protocol.

5. The distributed power supply power coordination optimization method based on edge computing according to claim 4 is characterized in that: The local correction of the output power and load pressure includes introducing the load difference, power difference and adjustment margin between the current node and the adjacent nodes, and dynamically adjusting the power output plan of each node based on the local collaborative optimization algorithm.

6. The distributed power supply power coordination optimization method based on edge computing according to claim 5, characterized in that: The current power grid fluctuation and economic signal data include power grid frequency, power grid interface voltage, real-time electricity price, and power purchase and feed-in power information.

7. The distributed power supply power coordination optimization method based on edge computing according to claim 6, characterized in that: The continuous monitoring of the power output effect and the predicted deviation includes real-time collection of the actual power output of each node, the deviation rate between the output and the power output plan, and triggering emergency compensation measures according to the deviation level.

8. A system using the distributed power supply power coordination optimization method based on edge computing according to any one of claims 1 to 7, characterized in that: It includes edge node state perception and short-cycle power prediction module, edge node collaborative optimization and local correction module, grid interactive regulation and closed-loop optimization feedback module; The edge node state perception and short-term power prediction module is used to collect operating parameters at each edge node in real time, build a short-term power prediction model based on the collected data, calculate future power generation capacity and load evolution trends, and output a local power output plan and preliminary pre-allocation value based on historical operating rules and load characteristics; The edge node collaborative optimization and local correction module is used to share the short-term predicted output value, operating load and remaining regulation capacity between edge nodes through a lightweight communication mechanism, and use the shared data to locally correct the power output plan of each node; The grid interactive control and closed-loop optimization feedback module is used to make the final coordinated optimization adjustment of the power output of each node, and continuously monitor the power output effect and prediction deviation during the scheduling execution phase.

9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the steps of the distributed power supply coordination optimization method based on edge computing according to any one of claims 1 to 7 are implemented.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the distributed power supply coordination optimization method based on edge computing according to any one of claims 1 to 7 are implemented.

Images