A photovoltaic power plant distributed energy collaborative control method and device

Abstract

The application discloses a kind of photovoltaic power plant distributed energy collaborative control method and device, it is related to intelligent power grid control technical field.This kind of photovoltaic power plant distributed energy collaborative control method and device, comprising: S1, collection inverter working data, and pre-processing is executed to inverter working data, obtain the inverter working data after pre-processing;S2, construct the inverter set to be classified, extract reference sample, assess the risk of withdrawal from network, output the withdrawal state label of inverter to be classified;S3, identify power intervention candidate area, extract inverter that has not withdrawn from network and construct power compensation resource set, screen out power regulation target, and assess the regulation priority of power regulation target;S4, calculate the power regulation step of power regulation target, carry out power adjustment, adjust after constructing feeder regulation database, realize the closed-loop adaptive optimization of power control.The problem that inverter withdrawal failure misjudgment caused by feeder end voltage drop anomaly not being identified is solved.

Description

Technical Field

[0001] This invention relates to the field of smart grid control technology, specifically to a method and device for coordinated control of distributed energy resources in photovoltaic power plants. Background Technology

[0002] In the grid-connected operation of photovoltaic power plants, the inverter, as a key conversion unit, is directly affected by the control strategy, which in turn impacts the grid connection quality and system stability. Existing technologies have proposed various control methods and operational analysis models to improve the grid connection performance and anomaly handling capabilities of inverters.

[0003] For example, invention CN111030173B discloses a control method and device for a grid-connected inverter in a new energy power plant, an inverter, and a dielectric. The control method for the grid-connected inverter in a new energy power plant includes: acquiring the inverter's input voltage measurement value, output voltage measurement value, and output current measurement value; wherein the input voltage measurement value is the bus voltage measurement value of the DC bus connected to the inverter; determining the inverter's active current setpoint based on the input voltage measurement value; determining the inverter's reactive current setpoint based on the output voltage measurement value and the output current measurement value; and generating a PWM signal using the active current setpoint and the reactive current setpoint, the PWM signal being used to control the operating state of the inverter's IGBTs so that the inverter's output voltage is close to the preset output voltage setpoint.

[0004] For example, the invention disclosed in CN117578405B, a method and device for analyzing anomalies in inverter strings of a photovoltaic power plant, includes: collecting historical operating data and design strategies of the inverter strings over a period of time; formulating capacity ratio rules based on the historical operating data and design strategies, and establishing a capacity ratio detection model based on the capacity ratio rules and historical operating data; collecting real-time operating data of the inverter strings, inputting the real-time operating data into the capacity ratio detection model, and determining whether the capacity ratio of the inverter strings is abnormal through the capacity ratio detection model. This invention establishes a capacity ratio detection model by collecting operating data and design strategies of the inverter strings to determine whether the capacity ratio of the inverter strings is abnormal in real time, thereby preventing further losses or failures caused by anomalies. This greatly improves monitoring efficiency and accuracy, ensures that the capacity ratio of the inverter strings can maintain a normal working state, and also improves the power generation efficiency and operational stability of the photovoltaic power plant.

[0005] However, although the above technologies have made some progress in inverter current and voltage control and string anomaly identification, they still have the following shortcomings: First, they fail to realize the inter-inverter status linkage judgment based on feeder topology, making it difficult to identify abnormal areas with coordinated grid disconnection characteristics; second, they lack quantitative modeling methods between inverter fault status and operating data, making it difficult to support targeted power regulation strategies; and third, they have not established a closed-loop feedback mechanism from status assessment to regulation execution, making it unable to adapt to real-time optimization control under dynamic grid conditions.

[0006] Therefore, in order to address the above problems, there is an urgent need for a method and device for the coordinated control of distributed energy in photovoltaic power plants. Summary of the Invention

[0007] Technical problems to be solved

[0008] To address the shortcomings of existing technologies, this invention provides a method and device for coordinated control of distributed energy in photovoltaic power plants, which solves the problem of misjudging inverter grid disconnection faults due to the failure to identify abnormal voltage drop at the end of the feeder.

[0009] Technical solution

[0010] To achieve the above objectives, the present invention provides the following technical solution: a method and apparatus for coordinated control of distributed energy in photovoltaic power plants, comprising: S1, collecting inverter operating data and performing cycle integrity verification, trend segmentation, anomaly suppression, and scale normalization on the inverter operating data to obtain preprocessed inverter operating data; S2, constructing a set of inverters to be classified based on de-grid inverters, extracting reference samples in conjunction with feeder topology, assessing de-grid risk based on preprocessed inverter operating data, outputting de-grid status labels of the inverters to be classified, and constructing an inverter de-grid dataset; S3, identifying power intervention candidate regions based on the de-grid dataset, extracting power compensation resource sets from non-de-grid inverters, screening power adjustment targets, and evaluating the adjustment priority of the power adjustment targets; S4, calculating the power adjustment step size of the power adjustment targets, combining the active power output command value after constraint verification with AC output power generation, performing power adjustment, and constructing a feeder adjustment database after adjustment to achieve closed-loop adaptive optimization of power control.

[0011] Furthermore, inverter operating data is collected, and the data undergoes cycle integrity verification, trend segmentation, anomaly suppression, and scale normalization. The specific steps for obtaining preprocessed inverter operating data are as follows: Based on a fixed time window, a sampling period is set, and inverter operating data from each photovoltaic inverter in the power plant is collected. This data includes inverter output voltage, inverter output current, bus current, bus voltage, ambient temperature, terminal box temperature, inverter disconnection times, grid connection response time, grid connection restart times, photovoltaic DC input power, and AC output power. Stability analysis is then performed to calculate... The method performs periodic integrity verification on inverter operating data, identifying and eliminating breakpoints and duplicate segments caused by communication interruptions, data backhaul delays, and momentary equipment disconnections; it performs multi-cycle steady-state segmentation processing on inverter operating data using a trend residual decomposition algorithm to separate natural fluctuations and short-term abnormal jumps during operation; it performs upper and lower limit suppression and data sparsity repair on inverter operating data through a grouped discrete anomaly suppression strategy to prevent low-frequency event values ​​from dominating regression results; and it performs range normalization processing on inverter operating data through an interval linear mapping function to map data with different physical dimensions to a unified standardized interval.

[0012] Furthermore, based on the decommissioned inverters, a set of inverters to be classified is constructed, and reference samples are extracted in conjunction with the feeder topology. The specific steps for assessing the decommissioning risk based on the preprocessed inverter operating data are as follows: Within the same sampling period, all inverter operating data are retrieved, inverters that have undergone decommissioning are screened, and a set of inverters to be classified is constructed; for each inverter to be classified, the feeder to which it belongs is identified according to the electrical access topology, and other non-decommissioned inverters on the feeder are extracted to construct a set of reference inverters, which serve as parallel comparison samples for the current inverters to be classified; based on the preprocessed inverter operating data, the risk of decommissioning is assessed. The risk of inverter de-grid conversion is calculated as follows: Divide the difference between the bus voltage and the inverter output voltage by the inverter output current to obtain the unit current voltage drop term; square the difference between the terminal box temperature and the ambient temperature to obtain the thermal disturbance temperature difference term; square the unit current voltage drop term and add it to the thermal disturbance temperature difference term to obtain the voltage drop and temperature difference combined disturbance term; divide the number of grid restarts by the grid response time, add one to the ratio and take the natural logarithm to obtain the restart response factor; multiply the voltage drop and temperature difference combined disturbance term by the restart response factor to obtain the load coupling excitation term; divide the load coupling excitation term by the sum of the photovoltaic DC input power and the minimum term to obtain the de-grid conversion risk assessment value.

[0013] Further, the specific steps for constructing the inverter decommissioning dataset by outputting the decommissioning status label of the inverter to be classified are as follows: Real-time comparison of the decommissioning risk assessment value and the decommissioning risk threshold. When the decommissioning risk assessment value of the inverter to be classified is less than or equal to the decommissioning risk threshold, the inverter to be classified is determined to be in a false alarm decommissioning state and is removed from the classification set. When the decommissioning risk assessment value of the inverter to be classified is greater than the decommissioning risk threshold, and there is a corresponding reference inverter whose decommissioning risk assessment value is greater than the decommissioning risk threshold, the inverter to be classified is determined to be in a line-type decommissioning state. Otherwise, the inverter to be classified is determined to be in an equipment-type decommissioning state. Output the decommissioning status label of the inverter to be classified, and perform structured storage with the corresponding inverter operating data to construct the inverter decommissioning dataset.

[0014] Furthermore, based on the grid decommissioning dataset, the specific steps for identifying power intervention candidate regions, extracting non-decommissioned inverters to construct a power compensation resource set, and selecting power regulation targets are as follows: Clustering of line-type grid decommissioning behavior is performed on the inverter grid decommissioning dataset of the same cycle, grouping by feeder number, and selecting regions at the end of the same feeder where there are line-type decommissioned inverters meeting a set quantity condition, marking them as power intervention candidate regions; within each power intervention candidate region, the inverters are sorted according to the feeder physical topology order, and the non-decommissioned inverters at the feeder starting point are extracted to form a power compensation resource set; in the power compensation resource set, inverters meeting the following conditions are selected as power regulation targets: the ratio of photovoltaic DC input power to AC output power is greater than the regulation threshold, and the inverter output current is less than the rated operating current.

[0015] Furthermore, the specific steps for evaluating the adjustment priority of the power regulation target are as follows: squaring the difference between the photovoltaic DC input power and the AC output power to obtain the power regulation feasibility quantity; multiplying the difference between the terminal box temperature and the ambient temperature as the temperature control deviation term with the power regulation feasibility quantity to obtain the thermal coupling regulation driving factor; and dividing the thermal coupling regulation driving factor by the product of the grid disconnection risk assessment value and the inverter output current to obtain the power regulation priority assessment value.

[0016] Furthermore, the specific steps for calculating the power regulation step size of the power regulation target are as follows: For each power regulation target, establish a load regulation mapping relationship based on the corresponding feeder structure, subtract the AC output power from the photovoltaic DC input power to obtain the power regulation surplus; subtract the ratio of the inverter output current to the rated operating current to obtain the current regulation coefficient; take the natural logarithm of the sum of the grid disconnection risk assessment value and the minima to obtain the risk regulation factor; multiply the power regulation surplus, the current regulation coefficient, and the risk regulation factor in sequence to obtain the output power regulation step size.

[0017] Furthermore, the specific steps for power adjustment, based on the active power output command value after constraint verification of AC output power generation, are as follows: Calculate the active power output command value of the power adjustment target using a linear amplitude adjustment algorithm, based on the output power adjustment step size and the real-time AC output power; perform constraint boundary verification between the active power output command value and the rated output power, and filter out over-limit command values; select power adjustment targets sequentially in descending order of power adjustment priority evaluation values ​​and issue corresponding active power output adjustment commands to perform power adjustment operations.

[0018] Furthermore, the specific steps for constructing a feeder regulation database after adjustment to achieve closed-loop adaptive optimization of power control are as follows: After the power adjustment is completed, monitor the changes in the grid disconnection behavior of the remaining inverters in the corresponding feeder and update the grid disconnection status labels of the remaining inverters; the output power adjustment step size, active power output command value, grid disconnection status labels of the corresponding inverters in the feeder before and after adjustment, and inverter working data are structurally associated and stored to construct a feeder regulation database, which is used to support the adaptive optimization and strategy backtracking of subsequent power adjustments.

[0019] The second aspect of this invention provides a distributed energy collaborative control device for photovoltaic power plants, comprising: a data acquisition and preprocessing module, a risk assessment and status identification module, a power regulation target determination module, and an active power regulation control module. The data acquisition and preprocessing module is used to acquire inverter operating data and perform cycle integrity verification, trend segmentation, anomaly suppression, and scale normalization on the inverter operating data to obtain preprocessed inverter operating data. The risk assessment and status identification module is used to construct a set of inverters to be classified based on de-grid inverters, and extract reference samples based on the feeder topology, based on the preprocessed data. The system assesses grid disconnection risks based on inverter operating data, outputs grid disconnection status labels for inverters to be classified, and constructs an inverter grid disconnection dataset. A power regulation target determination module identifies candidate regions for power intervention based on the grid disconnection dataset, extracts power compensation resource sets from non-grid disconnected inverters, filters out power regulation targets, and evaluates the regulation priority of the power regulation targets. An active power regulation control module calculates the power regulation step size of the power regulation targets, combines the AC output power to generate active power output command values ​​after constraint verification, performs power adjustment, and constructs a feeder regulation database after adjustment to achieve closed-loop adaptive optimization of power control.

[0020] Beneficial effects

[0021] The present invention has the following beneficial effects:

[0022] (1) This method and device for distributed energy collaborative control in photovoltaic power plants constructs a dynamic power regulation mechanism for inverters, integrating multi-dimensional indicators such as DC input power, AC output power, current margin, and grid disconnection risk to achieve dynamic calculation and differentiated execution of power regulation step size, thereby optimizing the local load distribution of feeders and improving the collaborative operation capability between inverters. It further avoids feeder overload and voltage fluctuation problems caused by load concentration, ensuring the overall grid operation safety.

[0023] (2) This method and device for coordinated control of distributed energy in photovoltaic power plants integrates grid decommissioning risk assessment and control strategy response to form a closed-loop control chain from assessment, identification, regulation to backtracking, realizing full-process automation from state perception to intelligent decision-making, and significantly improving the system's intelligent regulation capability and adaptive level. It effectively addresses the regulation challenges brought about by changes in the operating environment and inverter state fluctuations, and has the ability to continuously evolve and adapt.

[0024] (3) The distributed energy collaborative control method and device for photovoltaic power plants extracts reference samples based on feeder topology and classifies out-of-grid behavior to achieve rapid identification of line-type out-of-grid and intelligent power yield control of front-end inverters, thereby effectively alleviating voltage drop disturbances at the end of feeders, reducing the probability of inverters being mistakenly out-of-grid, enhancing the power supply stability of the power plant, and avoiding misjudgment and misreplacement of the equipment itself, reducing operation and maintenance costs, and improving the accuracy and timeliness of fault location.

[0025] (4) This method and device for coordinated control of distributed energy in photovoltaic power plants extracts non-deregistered inverters to form a candidate set for regulation, and combines it with a regulation priority evaluation strategy to screen power targets with regulation potential, thereby achieving orderly and controllable local power redistribution and improving the utilization efficiency of photovoltaic system regulation resources. Under the premise of ensuring operational reliability, it maximizes the release of remaining adjustable capacity and achieves optimal resource scheduling. Attached Figure Description

[0026] Figure 1 This is a flowchart of a distributed energy collaborative control method for photovoltaic power plants;

[0027] Figure 2 This is a structural diagram of a distributed energy collaborative control device for a photovoltaic power plant;

[0028] Figure 3 Image showing the grid disconnection identification results for inverters to be classified;

[0029] Figure 4 Output power regulation pattern for power regulation target. Detailed Implementation

[0030] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0031] Please see Figures 1-4 This invention provides a technical solution: a method and device for coordinated control of distributed energy in a photovoltaic power plant, comprising: S1, collecting inverter operating data and performing cycle integrity verification, trend segmentation, anomaly suppression, and scale normalization on the inverter operating data to obtain preprocessed inverter operating data; S2, constructing a set of inverters to be classified based on de-grid inverters, extracting reference samples in conjunction with feeder topology, assessing de-grid risk based on preprocessed inverter operating data, outputting de-grid status labels for the inverters to be classified, and constructing an inverter de-grid dataset; S3, identifying power intervention candidate regions based on the de-grid dataset, extracting power compensation resource sets from non-de-grid inverters, screening power adjustment targets, and evaluating the adjustment priority of the power adjustment targets; S4, calculating the power adjustment step size of the power adjustment targets, combining the active power output command value after constraint verification of AC output power generation, performing power adjustment, and constructing a feeder adjustment database after adjustment to achieve closed-loop adaptive optimization of power control.

[0032] Specifically, the steps for collecting inverter operating data and performing cycle integrity verification, trend segmentation, anomaly suppression, and scale normalization on the inverter operating data to obtain preprocessed inverter operating data are as follows: Based on a fixed time window, a sampling period is set, and inverter operating data of each photovoltaic inverter in the power plant is collected. The inverter operating data consists of field operating parameters collected according to a unified time sequence, specifically including inverter output voltage, inverter output current, bus current, bus voltage, ambient temperature, terminal box temperature, number of times the inverter is disconnected from the grid, grid connection response time, number of times the grid connection is restarted, photovoltaic DC input power, and AC output power. Each data item has continuity and sampling time label to ensure the timeliness and traceability of data quality. A sampling stability analysis algorithm is used to verify the periodic integrity of the inverter operating data. This verification includes analyzing the periodic continuity of the inverter output voltage and current item by item, performing gap scanning on the time trajectories of bus voltage and current, identifying and removing breakpoints and repetitive segments caused by communication interruptions, data transmission delays, and momentary power outages, ensuring that subsequent processing is based only on high-confidence, valid data. A trend residual decomposition algorithm is used to perform multi-cycle steady-state segmentation of the inverter operating data, identifying and dividing the operating cycle containing the stable trends of photovoltaic DC input power and AC output power. During the decomposition process, true fluctuation characteristics are preserved while short-term abnormal disturbances are removed, effectively separating natural operating patterns from discrete anomalies, improving the representativeness and accuracy of the analysis. A grouped discrete anomaly suppression strategy is used to perform upper and lower limit suppression and data sparsity repair on the inverter operating data, preventing local extreme values ​​from causing imbalances in subsequent calculations. By performing range normalization processing on inverter operating data through an interval linear mapping function, operating data with different physical dimensions are mapped to a standard unified interval, keeping inverter operating data within a unified scale for comparison and fusion, and enhancing the correlation expression ability between different indicators.

[0033] In this implementation scheme, inverter operating data such as output voltage, output current, bus voltage, bus current, ambient temperature, terminal box temperature, number of inverter grid disconnections, grid connection response time, number of grid connection restarts, photovoltaic DC input power, and AC output power are acquired item by item. Periodic integrity verification, trend segmentation, anomaly suppression, and scale normalization are then performed, effectively improving the continuity, stability, representativeness, and standardization of the collected data. The resulting high-quality preprocessed dataset provides accurate and reliable support for subsequent grid disconnection risk identification, power regulation target selection, and active power output command generation, enhancing the adaptability and control accuracy of this invention under different operating conditions.

[0034] Specifically, the following steps are taken to assess the grid decommissioning risk based on the pre-processed inverter operating data: First, construct a set of inverters to be classified based on decommissioned inverters, and extract reference samples using feeder topology. Then, within the same sampling period, retrieve all inverter operating data, filter out inverters that have undergone grid decommissioning, and construct a set of inverters to be classified. For each inverter to be classified, determine the feeder number according to the electrical access topology, and extract all inverters that have not undergone grid decommissioning from their respective feeders to construct a reference inverter set. This set serves as a parallel comparison sample for the inverters to be classified, ensuring that the reference samples have a consistent feeder structure. Finally, based on the pre-processed inverter operating data after completing periodic integrity verification, trend segmentation, anomaly suppression, and scale normalization, execute the calculation process for the grid decommissioning risk assessment value item by item: First, divide the difference between the bus voltage and the inverter output voltage by the inverter output current. The following steps are taken: First, a unit current voltage drop term is obtained to characterize the voltage drop intensity under current load. Then, the difference between the terminal box temperature and the ambient temperature is squared to obtain a thermal disturbance temperature difference term, which reflects the temperature rise amplitude. The squared unit current voltage drop term is then summed with the thermal disturbance temperature difference term to obtain a voltage drop and temperature difference combined disturbance term, which describes the degree of transient coupling interference under combined effects. Next, the number of grid restarts is divided by the grid response time, and then one is added and the natural logarithm is taken to obtain the restart response factor, which reflects the response efficiency of the inverter during the recovery process after grid decommissioning. The voltage drop and temperature difference combined disturbance term is multiplied by the restart response factor to obtain a load coupling excitation term, which comprehensively measures the stability vulnerability of the inverter under the superposition of electrical and thermal disturbances. Finally, the load coupling excitation term is divided by the sum of the photovoltaic DC input power and the minimum term to calculate the grid decommissioning risk assessment value, which is used to achieve quantitative identification and fine differentiation of the grid decommissioning status. Among them, the minterms are small but non-zero positive real numbers used to avoid numerical instability caused by division by zero during calculations, and their range is [value range missing]. arrive Unless otherwise specified, all subsequent minima will adopt this definition and value range. The formula for constructing the above-mentioned grid decommissioning risk assessment value aims to comprehensively quantify the impact of three types of indicators—voltage drop intensity, thermal disturbance degree, and restart response efficiency—on the inverter's operational stability under the feeder topology. The unit current voltage drop term in the numerator reflects the voltage drop intensity "per unit current" at the inverter's AC output terminal. Combined with the thermal disturbance temperature difference term and the restart response factor, it constitutes the load coupling exacerbation term, reflecting the comprehensive exacerbation trend under the superposition of electrical and thermal disturbances. The photovoltaic DC input power introduced in the denominator is used to measure the proportion of power carried by the inverter in the feeder. The larger the value, the higher its operating intensity in the current feeder, that is, the greater its regulation potential; under the same coupling disturbance conditions, its grid decommissioning risk is relatively low. This assessment value not only reflects the inverter's current disturbance stress response state, but also forms a normalized measure of risk by comparing it with the input power, providing a comparable basis for grid decommissioning prediction and strategy formulation at the feeder level.

[0035] The specific formula for calculating the risk assessment value of network decommissioning is as follows:

[0036] ;

[0037] In the formula, S represents the risk assessment value for network decommissioning. Indicates bus voltage. Indicates the inverter output voltage. This indicates the inverter output current. Indicates the terminal box temperature. Indicates ambient temperature. Indicates the number of times the grid has been restarted. Indicates the grid connection response time. Indicates the DC input power of the photovoltaic system. Indicates a minus term.

[0038] In this embodiment, Table 1 is a data table of grid decommissioning risk assessment values, listing various operating parameters of the five inverters to be classified during grid-connected operation, including bus voltage, inverter output voltage, inverter output current, terminal box temperature, ambient temperature, number of grid-connected restarts, grid-connected response time, photovoltaic DC input power, and grid decommissioning risk assessment value. Specifically, for inverter F1, the bus voltage is 397.3 ohms, the inverter output voltage is 389.6 ohms, the inverter output current is 18.2 ohms, the terminal box temperature is 48.2 ohms, the ambient temperature is 34.5 ohms, the number of grid-connected restarts is 2, the grid-connected response time is 5.4 seconds, the photovoltaic DC input power is 7200 kW, and the grid decommissioning risk assessment value is 0.00822. In inverter F2, the bus voltage is 392.1V, the inverter output voltage is 385.2V, the inverter output current is 17.9N / m³, the terminal box temperature is 47.8°C, the ambient temperature is 34.0°C, the grid connection restart count is 3, the grid connection response time is 6.0 seconds, the photovoltaic DC input power is 7150W, and the grid disconnection risk assessment value is 0.01081. In inverter F3, the bus voltage is 399.5V, the inverter output voltage is 390.3V, the inverter output current is 18.5W / m³, the terminal box temperature is 49.0°C, the ambient temperature is 35.0°C, the grid connection restart count is 2, the grid connection response time is 5.0 seconds, the photovoltaic DC input power is 7250W, and the grid disconnection risk assessment value is 0.00911. In inverter F4, the bus voltage is 390.4 ohms, the inverter output voltage is 382.0 ohms, the inverter output current is 17.5 ohms, the terminal box temperature is 47.0 ohms, the ambient temperature is 33.5 ohms, the number of grid-connected restarts is 4, the grid-connected response time is 6.5 seconds, the photovoltaic DC input power is 7100 kW, and the de-grid decommissioning risk assessment value is 0.01233. In inverter F5, the bus voltage is 393.8 ohms, the inverter output voltage is 386.1 ohms, the inverter output current is 18.1 ohms, the terminal box temperature is 48.5 ohms, the ambient temperature is 34.8 ohms, the number of grid-connected restarts is 3, the grid-connected response time is 5.8 seconds, the photovoltaic DC input power is 7180 kW, and the de-grid decommissioning risk assessment value is 0.01091.

[0039] Table 1. Data Table of Risk Assessment Values ​​for Network Decommissioning

[0040]

[0041] like Figure 3The figure shows the grid disconnection risk assessment values ​​of five inverters to be classified during the collaborative control judgment process, and indicates whether grid disconnection behavior has occurred. The figure uses a bar chart to represent the grid disconnection risk assessment value of each inverter, with different colors indicating the judgment results: red indicates identification as grid disconnection status, and green indicates identification as not disconnection status. The corresponding assessment value is labeled above each bar for easy comparison. Specifically, the grid disconnection risk assessment values ​​of inverters F2, F4, and F5 are 0.01081, 0.01233, and 0.01091, respectively, exceeding the grid disconnection risk threshold and thus judged as grid disconnection status; the assessment values ​​of inverters F1 and F3 are 0.00822 and 0.00911, respectively, below the grid disconnection risk threshold and thus judged as false alarms of grid disconnection status.

[0042] In this implementation plan, by retrieving all inverter operating data within the same sampling period, a set of inverters to be classified and a set of reference inverters are constructed. Samples are then paired based on feeder topology relationships to form a highly consistent grid shutdown comparison system, enabling accurate classification of the causes of inverter grid shutdown behavior. Simultaneously, by utilizing the step-by-step calculation process of unit current voltage drop, thermal disturbance temperature difference, voltage drop and temperature difference combined disturbance, restart response factor, load coupling intensification, and grid shutdown risk assessment value, the coupling intensification effect of multi-dimensional operating conditions on grid shutdown behavior is comprehensively evaluated. This improves the ability to identify line-type and equipment-type grid shutdowns, effectively avoids misdiagnosis and misjudgment, and enhances the accuracy and robustness of inverter grid shutdown identification.

[0043] Specifically, the steps for outputting the grid decommissioning status label of the inverters to be classified and constructing the inverter grid decommissioning dataset are as follows: Real-time comparison of the grid decommissioning risk assessment value and the grid decommissioning risk threshold. When the grid decommissioning risk assessment value of the inverter to be classified is less than or equal to the grid decommissioning risk threshold, combined with the changing trends of indicators such as inverter output voltage, inverter output current, and grid connection response time, it is determined that the inverter to be classified is in a false alarm grid decommissioning state and is removed from the classification set to avoid the possibility of erroneous intervention. When the grid decommissioning risk assessment value of the inverter to be classified is greater than the grid decommissioning risk threshold, and there is a corresponding reference inverter whose grid decommissioning risk assessment value is also greater than the grid decommissioning risk threshold, and its corresponding feeder number is consistent with that of the inverter to be classified, it is determined that the inverter to be classified is in a line-type grid decommissioning state caused by feeder conduction, identifying it as a regional problem under topology influence. Otherwise, under the premise that no other reference inverters show a high grid decommissioning risk assessment value, combined with the changing characteristics of independent factors such as terminal box temperature, ambient temperature, and grid restart count, it is determined that the inverter to be classified is in an equipment-type grid decommissioning state caused by its own abnormality. Finally, the grid decommissioning status label of each inverter to be classified is output, and the grid decommissioning status label is structured and associated with inverter operating data such as inverter output voltage, inverter output current, photovoltaic DC input power, AC output power, bus voltage, bus current, terminal box temperature, ambient temperature, grid connection response time, and number of grid connection restarts to construct an inverter grid decommissioning dataset, which is used to support subsequent power regulation and status backtracking analysis.

[0044] In this implementation plan, by dynamically comparing the grid disconnection risk assessment value with the grid disconnection risk threshold, and combining multi-source inverter operating data such as inverter output voltage, inverter output current, photovoltaic DC input power, AC output power, terminal box temperature, ambient temperature, grid connection response time, and grid connection restart count, the system distinguishes between false grid disconnection status, line-type grid disconnection status, and equipment-type grid disconnection status, and outputs grid disconnection status labels to construct an inverter grid disconnection dataset. This enables accurate classification and structured storage of inverter operating status, effectively improving the accuracy of grid disconnection behavior identification and enhancing the targeting of power control strategies.

[0045] Specifically, the steps for identifying power intervention candidate regions based on the grid decommissioning dataset, extracting power compensation resource sets from non-decommissioned inverters, and selecting power regulation targets are as follows: Perform line-type grid decommissioning behavior clustering on the inverter grid decommissioning dataset within the same sampling period, grouping by feeder number, and selecting regions where there are line-type decommissioned inverters meeting a set quantity condition at the end of the same feeder. These regions are then marked as power intervention candidate regions based on the inverter decommissioning status label. Within each power intervention candidate region, all inverters are physically sorted according to the feeder's physical access topology, starting from the feeder's starting point. Inverters whose status is not yet disconnected from the grid are extracted to form a power compensation resource set. Among them, the feeder start side refers to the inverter connection position near the combiner box, and the feeder end side refers to the inverter connection position near the far end of the line, which is specifically identified according to the electrical topology diagram. In the power compensation resource set, based on the judgment condition that the ratio of photovoltaic DC input power to AC output power is greater than the adjustment threshold, and combined with the constraint that the inverter output current is less than the rated operating current, inverters with power regulation capability and potential are selected as the power regulation targets for this cycle, providing the basic regulation objects for the subsequent issuance of active power output control commands.

[0046] In this implementation scheme, line-type grid decommissioning behavior clustering is performed based on the inverter decommissioning dataset. Combined with feeder number and feeder physical topology order, candidate regions for power intervention are identified. Power compensation resource set is constructed by extracting non-decommissioned inverters at the feeder starting point. Then, by combining the ratio of photovoltaic DC input power to AC output power and the limitation that the inverter output current is less than the rated operating current, inverters with regulation capabilities are selected as power regulation targets. This achieves efficient identification and reasonable allocation of regulation resources, and improves the adaptive regulation capability of the photovoltaic system under local disturbances and the stability of feeder operation.

[0047] Specifically, the steps for evaluating the adjustment priority of power regulation targets are as follows: First, based on the current cycle's photovoltaic DC input power and AC output power, calculate the difference between the two and perform a square operation on the difference to obtain a power regulation feasibility quantity, which is used to measure the remaining space for the current inverter to perform active power regulation operations; then, extract the terminal box temperature and ambient temperature of the corresponding cycle, calculate the difference between the two as a temperature control deviation term, reflecting the level of thermal disturbance in the environment where the inverter is located; multiply the temperature control deviation term and the power regulation feasibility quantity to obtain a thermal coupling regulation driving factor, which is used to comprehensively evaluate the response strength of power regulation capability under changes in the thermal environment; finally, use the thermal coupling regulation driving factor as a regulation capability index, divide it by the product of the grid disconnection risk assessment value and the inverter output current. The grid disconnection risk assessment value reflects the stability of the feeder operation state where the inverter is located, and the inverter output current reflects the current load level. By multiplying the two, a stability constraint factor is constructed, and finally, a power regulation priority assessment value is obtained, which serves as the basis for ranking the execution order of power regulation targets.

[0048] The specific formula for calculating the power regulation priority evaluation value is as follows:

[0049] ;

[0050] In the formula, Q represents the priority evaluation value for power regulation. Indicates the DC input power of the photovoltaic system. Indicates AC output power. Indicates the terminal box temperature. The ambient temperature is represented by S, and the risk assessment value for network decommissioning is represented by S. This indicates the inverter output current.

[0051] In this implementation plan, after screening power regulation targets, key operating parameters such as photovoltaic DC input power, AC output power, terminal box temperature, ambient temperature, grid decommissioning risk assessment value, and inverter output current are introduced to establish a multi-parameter calculation mechanism for power regulation priority assessment value, enabling differentiated ranking control of power regulation targets. While ensuring that the power regulation feasibility quantity reflects the inverter's power redundancy, a thermally coupled regulation drive factor is introduced to comprehensively consider temperature control deviation characteristics. Simultaneously, the grid decommissioning risk assessment value and inverter output current are superimposed to form a stability constraint factor, thereby achieving a refined assessment of power regulation priority. This effectively improves the reliability judgment level and response regulation efficiency during the power regulation process, and enhances the collaborative management and control capabilities against load disturbances at the feeder end.

[0052] Specifically, the steps for calculating the power regulation step size of the power regulation target are as follows: For each power regulation target, establish a load regulation mapping relationship corresponding to the physical topology of the feeder based on the feeder structure to ensure that the power regulation behavior is consistent with the feeder electrical transmission path; subtract the AC output power from the photovoltaic DC input power to calculate the remaining power regulation capacity of the inverter; where the photovoltaic DC input power is the basis of the remaining power regulation capacity, which determines the amount of power that needs to be regulated. Use the ratio of the inverter output current to the rated operating current as a current regulation coefficient to characterize the current operating margin of the inverter; sum the grid disconnection risk assessment value with the set minimum term to avoid mathematical anomalies caused by the denominator being zero, and then take the natural logarithm of the sum to construct a risk regulation factor with nonlinear response characteristics, which is used to reflect the influence intensity of the inverter's current operating state on the regulation behavior; calculate the remaining power regulation capacity, the current regulation coefficient, and the risk regulation factor in a multiplicative order to finally obtain the output power regulation step size that comprehensively considers power redundancy, current margin, and grid disconnection stability.

[0053] The specific formula for calculating the output power adjustment step size is as follows:

[0054] ;

[0055] In the formula, Indicates the output power adjustment step size. Indicates the DC input power of the photovoltaic system. Indicates AC output power. This indicates the inverter output current. This represents the rated operating current, and S represents the risk assessment value for grid decommissioning. Indicates a minus term.

[0056] In this embodiment, Table 2 is a data table of output power adjustment step size, listing the key control parameter values ​​for five power adjustment targets in the distributed energy control process, including photovoltaic DC input power, inverter AC output power, inverter output current, rated operating current, grid disconnection risk assessment value, and the finally calculated output power adjustment step size. Specifically, in power adjustment target W1, the photovoltaic input power is 5200, the output power is 5100, the output current is 15.0, the rated current is 18.0, the grid disconnection risk assessment value is 0.015, and the calculated output power adjustment step size is −69.99; in W2, the photovoltaic input power is 5100, the output power is 5000, the output current is 16.2, the rated current is 18.0, the grid disconnection risk assessment value is 0.018, and the corresponding adjustment step size is −40.17; in W3, the photovoltaic input power is 5050, and the output power is 5080. The output current is 14.8 kW, the rated current is 18.0 kW, the grid disconnection risk assessment value is 0.009, and the corresponding adjustment step size is 25.12. In W4, the photovoltaic input power is 5300 kW, the output power is 5250 kW, the output current is 15.5 kW, the rated current is 18.0 kW, the grid disconnection risk assessment value is 0.013, and the adjustment step size is −30.16. In W5, the photovoltaic input power is 5150 kW, the output power is 5140 kW, the output current is 15.8 kW, the rated current is 18.0 kW, the grid disconnection risk assessment value is 0.017, and the adjustment step size is −4.98.

[0057] Table 2 Output Power Adjustment Step Size Data Table

[0058]

[0059] like Figure 4 The figure shows the output power adjustment step size of five power adjustment targets in the distributed cooperative control process. The figure uses a bar chart to indicate the output power adjustment step size of each power adjustment target. The direction of the bar indicates the adjustment trend: positive values ​​upwards indicate increased power, and negative values ​​downwards indicate decreased power. The bar color distinguishes the adjustment polarity, with blue indicating increased power and orange indicating decreased power. As can be seen from the figure, W3 is currently the only target with increased power, while the other power adjustment targets all decrease power. Figure 4 This intuitively demonstrates the fine adjustment capability of the present invention to achieve differentiated power control based on multi-factor comprehensive evaluation.

[0060] This implementation plan incorporates photovoltaic DC input power, AC output power, inverter output current, rated operating current, grid decommissioning risk assessment value, and minimum terms to construct power regulation margin, current regulation coefficient, and risk regulation factor. This forms a load regulation mapping mechanism consistent with the feeder structure, enabling dynamic calculation of the output power regulation step size. By fully integrating the inverter's real-time operating status, physical connection characteristics, and grid decommissioning stability indicators, it enhances the intelligent adaptation capability of power regulation targets while ensuring the accuracy of regulation response, thereby improving the flexibility of differentiated load regulation and the reliability of collaborative control.

[0061] Specifically, the power adjustment steps, based on the active power output command value after constraint verification of AC output power generation, are as follows: According to the output power adjustment step size and combined with the real-time acquired AC output power, a linear amplitude adjustment algorithm is used to calculate the active power adjustment amount for the power adjustment target, generating the corresponding active power output command value; the calculated active power output command value is compared item by item with the rated output power, and upper and lower limit boundary verification is performed to eliminate invalid command values ​​exceeding the rated output power range, ensuring that the active power output command value is within the executable range; based on the evaluated power adjustment priority assessment value, power adjustment targets are selected sequentially in descending order of value, and combined with the feeder number and physical topology location, the active power output adjustment command after boundary verification is issued, completing the AC output power adjustment operation for each target, improving the orderliness and differentiated accuracy of power adjustment within the photovoltaic power plant.

[0062] In this implementation plan, by combining the active power output command value after AC output power generation constraint verification for power adjustment, the dynamic linkage between the output power adjustment step size and the real-time AC output power is effectively realized. This ensures that the active power output command value can be accurately generated within the rated output power boundary range, and the stability of command calculation is improved through a linear amplitude adjustment algorithm. On this basis, combined with the sorting mechanism of power adjustment priority evaluation value, adjustment commands are issued in an orderly manner to ensure that the power adjustment target is executed step by step according to priority, avoiding adjustment conflicts and enhancing the controllability and practicality of active power output adjustment commands. This improves the execution accuracy and adjustment stability of the AC output power adjustment process within the photovoltaic power plant.

[0063] Specifically, the following steps are taken to construct a feeder regulation database after adjustment to achieve closed-loop adaptive optimization of power control: After power adjustment, continuously monitor the number of inverter disconnections and grid connection response times of the other inverters in the corresponding feeder. Identify the trend of disconnection status changes by observing changes in disconnection behavior. Update the disconnection status labels of the other inverters in real time based on the identification results and verify the stability and consistency of the label updates. Structure and associate the output power adjustment step size, active power output command value, disconnection status labels before and after adjustment of the corresponding inverters in the feeder, inverter output voltage, inverter output current, bus current, bus voltage, terminal box temperature, ambient temperature, number of grid connection restarts, photovoltaic DC input power, and AC output power of each power regulation target. Construct a feeder regulation database indexed by feeder number and sampling period. This database supports the generation of adaptive optimization regulation strategies based on the differences between historical power adjustment step sizes and disconnection status labels, and provides a data support basis for updating regulation strategies based on inverter working data sequence backtracking, thereby achieving closed-loop adaptive optimization of the power control loop.

[0064] In this implementation scheme, a feeder regulation database indexed by feeder number and sampling period is constructed. This database stores the output power regulation step size, active power output command value, pre-adjustment grid disconnection status label, post-adjustment grid disconnection status label, inverter output voltage, inverter output current, bus current, bus voltage, terminal box temperature, ambient temperature, grid restart count, photovoltaic DC input power, and AC output power in a structured manner. The grid disconnection status label is continuously updated after power adjustment, supporting the generation of adaptive optimization regulation strategies. This enhances the quantitative feedback capability of power adjustment effects, improves the closed-loop response accuracy of the power control process, and enhances the historical backtracking capability of the regulation strategy, thereby realizing dynamic power control adaptive optimization based on inverter operating data.

[0065] like Figure 2As shown, the second aspect of the present invention provides a distributed energy collaborative control device for photovoltaic power plants, comprising: a data acquisition and preprocessing module, a risk assessment and status identification module, a power regulation target determination module, and an active power regulation control module, wherein: the data acquisition and preprocessing module is used to acquire inverter operating data and perform cycle integrity verification, trend segmentation, anomaly suppression, and scale normalization processing on the inverter operating data to obtain preprocessed inverter operating data; the risk assessment and status identification module is used to construct a set of inverters to be classified based on the de-grid inverters, and extract reference samples in combination with the feeder topology, based on the preprocessed... The system assesses grid disconnection risks based on processed inverter operating data, outputs grid disconnection status labels for inverters to be classified, and constructs an inverter grid disconnection dataset. A power regulation target determination module identifies candidate regions for power intervention based on the grid disconnection dataset, extracts power compensation resource sets from non-grid disconnected inverters, filters out power regulation targets, and evaluates the regulation priority of the power regulation targets. An active power regulation control module calculates the power regulation step size of the power regulation targets, combines the AC output power to generate active power output command values ​​after constraint verification, performs power adjustment, and constructs a feeder regulation database after adjustment to achieve closed-loop adaptive optimization of power control.

[0066] This implementation plan constructs a distributed energy collaborative control device for photovoltaic power plants, comprising a data acquisition and preprocessing module, a risk assessment and status identification module, a power regulation target determination module, and an active power regulation control module. This device achieves high-completeness preprocessing of inverter operating data, including inverter output voltage, inverter output current, bus current, bus voltage, ambient temperature, terminal box temperature, inverter de-grid times, grid connection response time, grid connection restart times, photovoltaic DC input power, and AC output power. Combined with feeder topology, it enables de-grid risk assessment and de-grid status label identification. Furthermore, based on the de-grid dataset, it filters power regulation targets with differentiated regulation capabilities, dynamically calculates the output power regulation step size, and generates constrained and verified active power output command values. By constructing a feeder regulation database, it supports adaptive optimization of subsequent power regulation strategies, significantly improving the photovoltaic power plant's inverter collaborative control capabilities, de-grid identification accuracy, and intelligent regulation level.

[0067] It should be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article, or apparatus.

[0068] The preferred embodiments of the present invention disclosed above are merely illustrative of the invention. These preferred embodiments do not exhaustively describe all details, nor do they limit the invention to the specific implementations described. Clearly, many modifications and variations can be made based on the content of this specification. This specification selects and specifically describes these embodiments to better explain the principles and practical applications of the invention, thereby enabling those skilled in the art to better understand and utilize the invention. The invention is limited only by the claims and their full scope and equivalents.

Claims

1. A method for coordinated control of distributed energy resources in a photovoltaic power plant, characterized in that, Includes the following steps: S1 collects inverter operating data and performs cycle integrity verification, trend segmentation, anomaly suppression and scale normalization on the inverter operating data to obtain preprocessed inverter operating data. S2: Construct a set of inverters to be classified based on the inverters that have been decommissioned from the grid, extract reference samples in combination with the feeder topology, assess the decommissioning risk based on the preprocessed inverter working data, output the decommissioning status label of the inverters to be classified, and construct the inverter decommissioning dataset. The specific steps for constructing a set of inverters to be classified based on decommissioned inverters and extracting reference samples in conjunction with the feeder topology are as follows: Within the same sampling period, all inverter operating data are retrieved, inverters that have been de-grid are screened, and a set of inverters to be classified is constructed. For each inverter to be classified, the feeder to which it belongs is identified according to the electrical access topology, and other inverters that have not been de-grid on the feeder to which it belongs are extracted to construct a set of reference inverters, which serve as parallel comparison samples for the current inverter to be classified. The specific steps for assessing the risk of grid decommissioning based on the preprocessed inverter operating data are as follows: Divide the difference between the bus voltage and the inverter output voltage by the inverter output current to obtain the unit current voltage drop term; Square the difference between the terminal box temperature and the ambient temperature to obtain the thermal disturbance temperature difference term. The voltage drop term is squared and added to the thermal disturbance temperature difference term to obtain the voltage drop and temperature difference joint disturbance term; the grid restart number is divided by the grid response time, the ratio is incremented by one and the natural logarithm is taken to obtain the restart response factor; the voltage drop and temperature difference joint disturbance term is multiplied by the restart response factor to obtain the load coupling excitation term; the load coupling excitation term is divided by the sum of the photovoltaic DC input power and the minimum term to obtain the grid disconnection risk assessment value; S3. Based on the grid decommissioning dataset, identify candidate regions for power intervention, extract non-grid decommissioned inverters to construct a power compensation resource set, screen out power regulation targets, and evaluate the regulation priority of the power regulation targets; S4 calculates the power adjustment step size of the power adjustment target, combines the AC output power to generate the active power output command value after constraint verification, performs power adjustment, and then constructs the feeder adjustment database to realize closed-loop adaptive optimization of power control.

2. The method for coordinated control of distributed energy resources in a photovoltaic power plant according to claim 1, characterized in that: The specific steps for collecting inverter operating data and performing cycle integrity verification, trend segmentation, anomaly suppression, and scale normalization on the inverter operating data to obtain preprocessed inverter operating data are as follows: Based on a fixed time window and a set sampling period, inverter operating data of each photovoltaic inverter in the power plant is collected. The inverter operating data includes inverter output voltage, inverter output current, bus current, bus voltage, ambient temperature, terminal box temperature, number of times the inverter is disconnected from the grid, grid connection response time, number of times the grid connection is restarted, photovoltaic DC input power, and AC output power. The inverter operating data is periodically integrity verified by a sampling stability analysis algorithm to identify and eliminate breakpoints and duplicate segments caused by communication interruptions, data backhaul delays, and momentary equipment disconnections; the inverter operating data is processed by a trend residual decomposition algorithm to perform multi-cycle steady-state segmentation to separate natural fluctuations and short-term abnormal jumps during operation. By employing a grouped discrete anomaly suppression strategy, upper and lower limit suppression and data sparsity repair are performed on the inverter operating data to prevent low-frequency event values ​​from dominating the regression results. Furthermore, by using an interval linear mapping function, the range of the inverter operating data is normalized to map data with different physical dimensions to a unified standardized interval.

3. The method for coordinated control of distributed energy resources in a photovoltaic power plant according to claim 1, characterized in that: The specific steps for constructing the inverter shutdown dataset by outputting the grid disconnection status labels of the inverters to be classified are as follows: The system compares the grid disconnection risk assessment value and the grid disconnection risk threshold in real time. When the grid disconnection risk assessment value of the inverter to be classified is less than or equal to the grid disconnection risk threshold, the inverter to be classified is determined to be in a false alarm grid disconnection state and is removed from the set to be classified. When the grid disconnection risk assessment value of the inverter to be classified is greater than the grid disconnection risk threshold, and there is a corresponding reference inverter whose grid disconnection risk assessment value is greater than the grid disconnection risk threshold, the inverter to be classified is determined to be in a line-type grid disconnection state. Otherwise, the inverter to be classified is determined to be in a state of equipment-type grid decommissioning; Output the grid disconnection status label of the inverter to be classified, and store it in a structured manner with the corresponding inverter working data to construct the inverter grid disconnection dataset.

4. The method for coordinated control of distributed energy resources in a photovoltaic power plant according to claim 1, characterized in that: The specific steps for identifying candidate regions for power intervention based on the grid disconnection dataset, extracting non-grid disconnected inverters to construct a power compensation resource set, and selecting power regulation targets are as follows: Clustering of line-type grid disconnection behavior is performed on the inverter grid disconnection dataset of the same cycle. The data is grouped by feeder number, and regions with line-type grid disconnection inverters that meet the set quantity conditions at the end of the same feeder are selected and marked as power intervention candidate regions. Within each power intervention candidate region, inverters are sorted by feeder physical topology, and the non-deactivated inverters at the feeder start point are extracted to form a power compensation resource set. In the power compensation resource set, inverters that meet the following conditions are selected as power regulation targets: the ratio of photovoltaic DC input power to AC output power is greater than the regulation threshold, and the inverter output current is less than the rated operating current.

5. The method for coordinated control of distributed energy resources in a photovoltaic power plant according to claim 1, characterized in that: The specific steps for evaluating the adjustment priority of the power adjustment target are as follows: The square of the difference between the photovoltaic DC input power and the AC output power is used to obtain the power regulation feasibility factor; the difference between the terminal box temperature and the ambient temperature is used as the temperature control deviation term, which is multiplied by the power regulation feasibility factor to obtain the thermal coupling regulation driving factor. Divide the thermal coupling regulation drive factor by the product of the grid disconnection risk assessment value and the inverter output current to obtain the power regulation priority assessment value.

6. The distributed energy collaborative control method for photovoltaic power plants according to claim 1, characterized in that: The specific steps for calculating the power adjustment step size of the power adjustment target are as follows: For each power regulation target, a load regulation mapping relationship is established based on the corresponding feeder structure. The power regulation surplus is obtained by subtracting the AC output power from the photovoltaic DC input power. The current regulation coefficient is obtained by subtracting the ratio of the inverter output current to the rated operating current. The risk regulation factor is obtained by taking the natural logarithm of the sum of the grid disconnection risk assessment value and the minimum term. The output power regulation step size is obtained by multiplying the remaining power regulation amount, the current regulation coefficient, and the risk regulation factor in sequence.

7. The method for coordinated control of distributed energy resources in a photovoltaic power plant according to claim 1, characterized in that: The specific steps for adjusting the power output by combining the AC output power with the constrained verification-generated active power output command value are as follows: Based on the output power adjustment step size and combined with the real-time AC output power, the active power output command value of the power adjustment target is calculated using a linear amplitude adjustment algorithm; the active power output command value and the rated output power are checked against the constraint boundary, and over-limit command values ​​are screened out; the power adjustment targets are selected in descending order of the power adjustment priority evaluation value and the corresponding active power output adjustment command is issued to perform the power adjustment operation.

8. The method for coordinated control of distributed energy resources in a photovoltaic power plant according to claim 1, characterized in that: The specific steps for constructing the feeder regulation database after adjustment to achieve closed-loop adaptive optimization of power control are as follows: After the power adjustment is completed, monitor the changes in the grid disconnection behavior of the remaining inverters in the corresponding feeder and update the grid disconnection status labels of the remaining inverters. The output power adjustment step size, active power output command value, grid disconnection status labels of the corresponding inverters in the feeder before and after adjustment and the inverter working data are structured and stored to build a feeder adjustment database to support the adaptive optimization and strategy backtracking of subsequent power adjustments.

9. A distributed energy collaborative control device for photovoltaic power plants, characterized in that: include: The module comprises a data acquisition and preprocessing module, a risk assessment and status identification module, a power regulation target determination module, and an active power regulation and control module, wherein: The data acquisition and preprocessing module is used to acquire inverter operating data and perform cycle integrity verification, trend segmentation, anomaly suppression and scale normalization on the inverter operating data to obtain preprocessed inverter operating data. The risk assessment and status identification module is used to construct a set of inverters to be classified based on the inverters that have been decommissioned, extract reference samples in combination with the feeder topology, assess the decommissioning risk based on the preprocessed inverter working data, output the decommissioning status label of the inverters to be classified, and construct the inverter decommissioning dataset. The specific steps for constructing a set of inverters to be classified based on decommissioned inverters and extracting reference samples in conjunction with the feeder topology are as follows: Within the same sampling period, all inverter operating data are retrieved, inverters that have been de-grid are screened, and a set of inverters to be classified is constructed. For each inverter to be classified, the feeder to which it belongs is identified according to the electrical access topology, and other inverters that have not been de-grid on the feeder to which it belongs are extracted to construct a set of reference inverters, which serve as parallel comparison samples for the current inverter to be classified. The specific steps for assessing grid decommissioning risk based on preprocessed inverter operating data are as follows: Divide the difference between the bus voltage and the inverter output voltage by the inverter output current to obtain a unit current voltage drop term; square the difference between the terminal box temperature and the ambient temperature to obtain a thermal disturbance temperature difference term; square the unit current voltage drop term and add it to the thermal disturbance temperature difference term to obtain a voltage drop and temperature difference combined disturbance term; divide the number of grid restarts by the grid response time, add one to the ratio, and take the natural logarithm to obtain a restart response factor; multiply the voltage drop and temperature difference combined disturbance term by the restart response factor to obtain a load coupling excitation term; divide the load coupling excitation term by the sum of the photovoltaic DC input power and the minimum term to obtain the grid decommissioning risk assessment value. The power regulation target determination module is used to identify candidate regions for power intervention based on the grid decommissioning dataset, extract power compensation resource sets from non-grid decommissioned inverters, screen out power regulation targets, and evaluate the regulation priority of the power regulation targets. The active power regulation and control module is used to calculate the power regulation step size of the power regulation target, generate an active power output command value after constraint verification based on the AC output power, perform power adjustment, and build a feeder regulation database after adjustment to realize closed-loop adaptive optimization of power control.

Images