Distributed photovoltaic fusion control method and device

Abstract

The invention relates to a distributed photovoltaic fusion control method and device, and relates to the field of photovoltaic control, and the method comprises the steps: obtaining equipment control parameters for state mapping when a photovoltaic station does not trigger an island effect, and obtaining an equipment operation baseline state; equipment state information is received according to a default period, if the equipment state information is consistent with the equipment operation baseline state, an equipment state fluctuation curve is constructed to perform same-mode high-frequency abnormal period indexing, and an equipment state monitoring period is obtained to update the default period; and when the photovoltaic station triggers an islanding effect or the equipment state information is inconsistent with the equipment operation baseline state, performing photovoltaic station off-grid control through a photovoltaic circuit breaker, thereby solving the technical problems of response lag and inflexible control caused by fixed period monitoring in photovoltaic management.

Description

Technical Field

[0001] The present invention relates to the field of photovoltaic control, and in particular to a fusion control method and device for distributed photovoltaic power generation. Background Art

[0002] With the increasing global emphasis on renewable energy and growing demand, distributed photovoltaic power generation, as a clean and efficient energy utilization method, has been widely applied. However, the effective management and monitoring of distributed photovoltaic sites have become an urgent problem to be solved.

[0003] Traditional distributed photovoltaic management usually relies on a fixed-period monitoring method, which has obvious lag and is difficult to respond promptly to abnormal situations in photovoltaic sites. For example, when an islanding effect occurs in a photovoltaic site or the operating state of equipment is abnormal, the traditional monitoring method cannot quickly detect and take control measures, resulting in problems such as equipment damage and power supply interruption. In addition, existing photovoltaic site monitoring and control schemes often lack intelligence and self-adaptability. They usually can only operate according to preset parameters and rules and cannot be flexibly adjusted, which limits the further improvement of the power generation efficiency and power supply stability of photovoltaic sites. Summary of the Invention

[0004] In view of the technical problems of lagging response and inflexible control caused by fixed-period monitoring in existing photovoltaic management, the present invention provides a fusion control method and device for distributed photovoltaic power generation to solve these problems.

[0005] The technical solution of the present invention to solve the above technical problems is as follows:

[0006] In a first aspect, the present invention provides a fusion control method for distributed photovoltaic power generation, which is applied to a master station. The master station is communicatively connected to a photovoltaic site. The method includes: when the photovoltaic site does not trigger an islanding effect, obtaining device control parameters for state mapping to obtain a baseline state of device operation; receiving device status information at a default period. If the device status information is consistent with the baseline state of device operation, constructing a device status fluctuation curve for co-modal high-frequency abnormal period indexing to obtain an updated default period for device status monitoring; when the photovoltaic site triggers an islanding effect or the device status information is inconsistent with the baseline state of device operation, performing off-grid control on the photovoltaic site through a photovoltaic circuit breaker.

[0007] In a second aspect, the present invention provides a fusion control device for distributed photovoltaic power generation, which is applied to a master station. The master station is communicatively connected to a photovoltaic site. The device includes: a parameter acquisition module, configured to obtain device control parameters for state mapping and obtain a baseline state of device operation when the photovoltaic site does not trigger an islanding effect; a state discrimination module, configured to receive device state information at a default period, and if the device state information is consistent with the baseline state of device operation, construct a device state fluctuation curve for co-modal high-frequency abnormal period indexing to obtain an updated default period of the device state monitoring period; and an off-grid control module, configured to perform off-grid control of the photovoltaic site through a photovoltaic circuit breaker when the photovoltaic site triggers an islanding effect or the device state information is inconsistent with the baseline state of device operation.

[0008] The beneficial effects of the present invention are as follows: Through dynamic period monitoring and intelligent control strategies, real-time, efficient, and precise management of photovoltaic sites is achieved, thereby ensuring the stability of photovoltaic power supply and quickly responding to site anomalies to avoid faults. BRIEF DESCRIPTION OF THE DRAWINGS

[0009] Figure 1 It is a schematic flowchart of a fusion control method for distributed photovoltaic power generation provided by the present invention.

[0010] Figure 2 It is a schematic structural diagram of a fusion control device for distributed photovoltaic power generation provided by the present invention.

[0011] Description of reference numerals: Parameter acquisition module 11, state discrimination module 12, off-grid control module 13. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0012] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative efforts belong to the scope of protection of the present invention.

[0013] In the description of the present invention, the terms "first" and "second" are only used for descriptive purposes and cannot be construed as indicating or implying relative importance or implicitly specifying the quantity of the indicated technical features. Thus, the features defined with "first" and "second" may explicitly or implicitly include one or more of the described features. In the description of the present invention, "a plurality" means two or more unless otherwise specifically defined.

[0014] In the description of the present invention, the term "for example" is used to mean "serving as an example, illustration, or explanation". Any embodiment described as "for example" in the present invention is not necessarily construed as being more preferred or advantageous than other embodiments. The following description is given to enable any person skilled in the art to implement and use the present invention. In the following description, details are set forth for purposes of explanation. It should be understood that those of ordinary skill in the art can recognize that the present invention can be implemented without using these specific details. In other instances, well-known structures and processes are not elaborated in detail to avoid obscuring the description of the present invention with unnecessary details. Therefore, the present invention is not intended to be limited to the embodiments shown, but is to be accorded the widest scope consistent with the principles and features disclosed herein.

[0015] Embodiment 1:

[0016] As Figure 1 shown, an embodiment of the present invention provides a fusion control method for distributed photovoltaic, which is applied to a master station. The master station is communicatively connected to photovoltaic sites. The method includes:

[0017] S10: When the photovoltaic site does not trigger the islanding effect, obtain device control parameters for state mapping to obtain the baseline state of device operation.

[0018] S20: Receive device status information according to a default period. If the device status information is consistent with the baseline state of device operation, construct a device status fluctuation curve for co-modal high-frequency anomaly period indexing to obtain an update of the default period of the device status monitoring period.

[0019] S30: When the photovoltaic site triggers the islanding effect, or when the device status information is inconsistent with the baseline state of device operation, perform off-grid control of the photovoltaic site through a photovoltaic circuit breaker.

[0020] Exemplarily, the master station is responsible for receiving data from each photovoltaic site, performing data analysis and processing, and sending control instructions to the photovoltaic site according to the analysis results to achieve efficient and stable operation of the entire distributed photovoltaic system. The photovoltaic site, that is, the installation location of the photovoltaic power generation system, is the basic unit in the distributed photovoltaic system. It mainly consists of components such as solar cell modules, storage batteries, controllers, and inverters, and can independently complete the collection, conversion, and storage of solar energy. The photovoltaic site converts solar energy into electrical energy through solar cell modules, and converts the electrical energy into alternating current suitable for use in the power grid through a controller and an inverter. At the same time, the photovoltaic site also has data collection and communication functions, and can upload its own operating status and data to the master station in real time for the master station to perform unified management and control. The control method of the present application is implemented by being applied to the master station and relying on the communication connection between the master station and the photovoltaic site.

[0021] Specifically, the islanding effect refers to the phenomenon that under certain conditions, after the photovoltaic power generation system is disconnected from the power grid, it still continues to supply power to independent loads, forming a self-sufficient power supply island. When the photovoltaic site does not trigger the islanding effect, it means that the connection between the photovoltaic power generation system and the power grid is normal, and the electric energy generated by the photovoltaic power generation can be smoothly transmitted to the power grid, or the electric energy can be obtained from the power grid to meet the load demand. At this time, the master station can send query instructions through the communication connection with the photovoltaic site to obtain the device control parameters of the current photovoltaic inverter. These parameters include limit control, fixed value control, proportional control, and timing control, which are used to control key indicators such as the output power and power factor of the photovoltaic inverter to ensure the stable operation of the photovoltaic site. Among them, the limit control means that the master station issues a limit control instruction to limit the active power and reactive power output of the photovoltaic inverter not to exceed the preset set value. This is to prevent the photovoltaic inverter from overloading or generating excessive reactive power, which may affect the stable operation of the power grid. The fixed value control means that the master station issues a fixed value control instruction to adjust the power factor of the photovoltaic inverter to the preset set value. This helps to maintain the power balance and voltage stability of the power grid. The proportional control means that the master station issues an inverter rated power ratio control instruction to control the active power and reactive power of the photovoltaic inverter not to exceed the set ratio value. This control method can flexibly adjust the output power according to the actual situation of the photovoltaic site to adapt to different lighting conditions and power grid requirements. The timing control means that the master station issues a timing control instruction, and the photovoltaic acquisition and monitoring unit controls the photovoltaic inverter in the way of limit control or proportional control according to the preset time period. This control method can realize the timed start-stop or power adjustment of the photovoltaic site to adapt to different operation requirements. After that, the master station maps the operation state of the photovoltaic inverter based on the obtained device control parameters and combined with the real-time operation data of the photovoltaic site, such as active power, reactive power, power factor, etc. The purpose of state mapping is to convert complex operation data into indicators that are easy to understand and analyze, so that the master station can monitor and manage the operation state of the photovoltaic site. Through state mapping, the master station can determine the baseline parameters of the photovoltaic inverter in the normal operation state. The device operation baseline parameters refer to a series of preset or actual operation parameters of the key devices, mainly the photovoltaic inverter, of the photovoltaic site in the normal operation state. These parameters constitute the benchmark or reference point for the operation of the photovoltaic site, which are used to evaluate the operation state of the device, conduct fault early warning, and optimize performance. These baseline parameters include, but are not limited to, the set values and actual operation values of active power, reactive power, power factor, etc. The establishment of the baseline state provides a basis for subsequent device monitoring, fault early warning, and performance optimization. To sum up, by obtaining device control parameters and performing state mapping, the master station can establish the device operation baseline state of the photovoltaic site, providing strong support for subsequent device monitoring, fault early warning, and performance optimization.

[0022] Further, in a device monitoring system, the system is set to receive status information from the device at a pre-determined default period, which is set manually and used to regularly check and update the device status information. These status information contain various operating parameters of the device and are used to reflect the current condition of the device. After that, whenever the system receives this status information, it will compare this information with the previously established baseline status of the device operation. The baseline status of the device operation is obtained based on historical data when the device is operating normally and represents the range of operating parameters of the device under ideal conditions. During the comparison process, the status parameters of each attribute (such as power, temperature, pressure, etc.) in the status information are compared one by one. If the status parameter of a certain attribute exceeds the interval set by the baseline status, and the proportion of this out-of-interval situation in the total time is greater than or equal to a pre-set time proportion threshold, then it is considered that this device is inconsistent with the baseline status. On the contrary, if the status parameters of all attributes do not exceed the baseline interval, or the proportion of the out-of-interval time is less than the time proportion threshold, it is regarded that the device is consistent with the baseline status. For example, suppose there is a power parameter of a device, and its baseline interval is 80% - 120% of the rated power. The pre-set time proportion threshold is 10%. Now, receive the device status information according to the default period (such as once a day). On a certain day, the received power parameter of this device is as follows: the proportion of time below 80% of the rated power is 5%, the proportion of time between 80% - 120% of the rated power is 85%, and the proportion of time above 120% of the rated power is 10%. Since the proportion of out-of-baseline interval time (10%) is equal to the time proportion threshold (10%), at this time, according to the rule that if the proportion of out-of-baseline interval time is greater than or equal to the time proportion threshold, it is regarded as inconsistent, that is, it is determined to be regarded as inconsistent with the baseline status. Next, if the device is consistent with the baseline status, the status fluctuation curve of the device will be further constructed. This curve reflects the change trend of the device status parameters over time. By analyzing this curve, the same-modal high-frequency abnormal periods in the device status can be identified, that is, those abnormal status periods that occur frequently and have similar patterns. Finally, based on the information of these abnormal periods, the status monitoring period of the device can be updated. This new period may be shorter or longer than the default period, depending on the frequency and pattern of the device status change. The updated period will be used as the new default period for the next reception and comparison of the device status information. For example, if it is found that the device status fluctuates frequently and has similar patterns, the system may shorten the monitoring period to once every half day or once every hour to capture and handle abnormal status more timely. On the contrary, if the device status is stable and there are few abnormalities, the system may extend the monitoring period to once every two days or once a week. To sum up, by setting the default period to receive the device status information, the current status of the device can be obtained regularly and automatically without manual intervention, thus improving the automation level of operation and maintenance.When the device is in line with the baseline state, the system does not need to take further actions, reducing unnecessary intervention and resource waste. When an anomaly is detected, the system can respond quickly and notify the operation and maintenance personnel in a timely manner for handling, shortening the time for fault detection and repair. By comparing with the baseline state of the device operation, it can accurately determine whether the device is in a normal state, avoiding operation and maintenance risks caused by false alarms or missed alarms. By analyzing the device status fluctuation curve, it can identify the same-mode high-frequency anomaly cycles in the device status, providing more accurate fault warning and location information for the operation and maintenance personnel. At the same time, it can dynamically adjust the monitoring period according to the change of the device status, avoiding resource waste and monitoring blind spots brought by fixed-period monitoring. When the device status is stable, extending the monitoring period can reduce the frequency of data transmission and processing, reducing system energy consumption and operation and maintenance costs. When the device status is abnormal, shortening the monitoring period can capture and handle the abnormal status more timely, improving the reliability and security of the system.

[0023] In a specific embodiment, when the photovoltaic site triggers the islanding effect, or the device status information is inconsistent with the baseline state of the device operation, this means that the photovoltaic site or the devices in it may encounter problems and immediate actions need to be taken to prevent potential safety risks. At this time, the system will perform off-grid control of the photovoltaic site through the photovoltaic circuit breaker according to the preset safety strategy. The photovoltaic circuit breaker is a switching device specially designed for photovoltaic systems, which can quickly cut off the circuit when detecting abnormal conditions to ensure the safe isolation of the photovoltaic site from the power grid.

[0024] The fusion control method for distributed photovoltaics provided by the embodiments of the present invention has at least the following technical effects:

[0025] 1. When the photovoltaic site does not trigger the islanding effect, by obtaining the device control parameters and performing state mapping, the baseline state of the device operation can be accurately established, which provides a reliable reference standard for subsequent anomaly detection. At the same time, real-time status monitoring ensures the timely update of the device status, helping to detect potential problems in a timely manner and avoiding the expansion of faults.

[0026] 2. When the device status is normal, by constructing the device status fluctuation curve and performing the same-mode high-frequency anomaly cycle indexing, the device status monitoring period can be intelligently adjusted. The intelligent cycle adjustment mechanism not only reduces unnecessary data transmission and processing, reduces system energy consumption, but also improves the accuracy and timeliness of anomaly detection.

[0027] 3. When the photovoltaic site triggers the islanding effect or the device status information is inconsistent with the baseline state of the device operation, the system can respond quickly and perform off-grid control of the photovoltaic site through the photovoltaic circuit breaker. This rapid response mechanism ensures the safe operation of the photovoltaic site under abnormal conditions, avoiding problems such as voltage and frequency instability caused by the islanding effect or device failures.

[0028] In a preferred embodiment, the obtaining of the device control parameters includes: when the islanding effect is not triggered at the photovoltaic site, when the first fluctuation value between the monitored grid load and the pre-stored grid load is greater than or equal to the first fluctuation threshold, and / or the second fluctuation value between the monitored environmental factors and the pre-stored environmental factors is greater than or equal to the second fluctuation threshold, inputting the monitored grid load and the monitored environmental factors into the historical control database to retrieve the device control parameters; using the monitored grid load to update the pre-stored grid load, using the monitored environmental factors to update the pre-stored environmental factors, and at the same time sending the device control parameters to the photovoltaic site for control initialization.

[0029] Specifically, when the system is not in the islanding effect state, it will dynamically adjust the device control parameters according to the fluctuations of the grid load and environmental factors. When the photovoltaic site is operating normally, it continuously monitors the grid load and environmental factors (such as light intensity, temperature, etc.), and compares this data with the pre-stored grid load and environmental factors. The pre-stored data here can be regarded as the reference value when the system was last operating stably. When the first fluctuation value between the monitored grid load and the pre-stored grid load reaches or exceeds the preset first fluctuation threshold, or the second fluctuation value between the monitored environmental factors and the pre-stored environmental factors reaches or exceeds the preset second fluctuation threshold, the system believes that a significant change has occurred in the current state and the control parameters need to be optimized. To obtain the optimal device control parameters, the current monitored grid load and monitored environmental factors are used as input conditions and retrieved from the historical control database. This database stores the control parameters of photovoltaic sites that have successfully operated in similar states in the past. Through the retrieval, the historical control parameters that best match the current state can be found, and these parameters are regarded as the optimal control strategy in the current state. Once the appropriate device control parameters are retrieved, these new parameters will be immediately used to update the pre-stored grid load and environmental factor data to ensure an updated reference value for the next comparison. At the same time, these new device control parameters will be sent to the photovoltaic site for initializing or adjusting the control strategy of the site. Whether the photovoltaic site triggers the islanding effect is an important prerequisite throughout this process. The islanding effect refers to the situation where when the power grid loses power, the photovoltaic power generation system continues to supply power to the de-energized grid line, forming a self-powered island that the power company cannot control. To detect the islanding effect, the photovoltaic site will adopt various methods, such as the active frequency shift detection method, the active phase shift detection method, and the voltage positive feedback type islanding effect detection method, etc. These methods judge whether the islanding effect occurs by applying small disturbances to the power grid and observing its response. Once the islanding effect is detected, corresponding protection measures will be immediately taken to ensure the safety of personnel and equipment. The specific detection methods are publicly available in the prior art and will not be elaborated here. In summary, this process is essentially an intelligent control strategy based on historical database retrieval. It can quickly retrieve the optimal device control parameters and make corresponding adjustments when the grid load and environmental factors fluctuate, thereby ensuring the stable operation and optimized control of the photovoltaic site. At the same time, by continuously monitoring and updating the pre-stored data, it can continuously adapt to changes in the external environment and improve the overall control efficiency and accuracy.

[0030] In a preferred embodiment, the retrieving of the device control parameters by inputting the grid load and environmental factors into the historical control database includes: inputting the grid load and the environmental factors into the historical control database to retrieve multiple groups of initial device control parameters of the photovoltaic site, wherein the multiple groups of initial device control parameters have multiple fitness identifiers, and the greater the fitness, the better the control parameters; using the multiple groups of initial device control parameters as the initial population, performing population iteration according to the multiple fitness identifiers to obtain updated device control parameters; taking the grid load, the environmental factors and the background conditions of the photovoltaic site layout topology configuration, and setting the updated device control parameters as the foreground conditions, and retrieving the power factor eigenvalue, the control stability eigenvalue and the dynamic response speed eigenvalue of the control samples that meet the background conditions and the foreground conditions through networking; processing the power factor eigenvalue, the control stability eigenvalue and the dynamic response speed eigenvalue through a fitness evaluation function to obtain an updated parameter fitness identifier; when the updated parameter fitness identifier is less than or equal to the minimum value of the multiple fitness identifiers, performing cyclic iteration based on the multiple groups of initial device control parameters; when the updated parameter fitness identifier is greater than the minimum value of the multiple fitness identifiers, using the updated device control parameters to replace the initial device control parameters with the minimum value of the fitness identifier of the multiple groups of initial device control parameters to obtain an updated population for cyclic iteration; when the number of cyclic iterations meets the preset number, outputting the device control parameters with the maximum fitness identifier.

[0031] Optionally, the real-time monitored grid load and environmental factors are used as input conditions and fed into the historical control database. This database stores a large number of device control parameters of photovoltaic sites under similar conditions in the past, and these parameters are all marked with fitness indicators. The larger the indicator, the better the effect of the control parameter. Multiple groups of initial device control parameters matching the current conditions can be retrieved from the database, and these parameters form the initial population for subsequent optimization. Then, the population is iteratively updated according to the fitness indicators of these parameters. During the iteration process, the parameters will be continuously adjusted in an attempt to obtain a better control effect. In order to more comprehensively evaluate the performance of the parameters, the current grid load, environmental factors, and the layout topology configuration of the photovoltaic site are used as background conditions, and the iteratively updated device control parameters are used as foreground conditions to retrieve control samples that meet these conditions through network connection. Among them, the background conditions refer to factors that are relatively stable, not easily changed, and have an important impact on the operation of the photovoltaic site. These factors are set as the basic conditions for retrieval when retrieving control samples through network connection, and are used to screen out historical data or cases similar to the current situation. The grid load included therein refers to the power demand borne by the grid at a certain moment. The magnitude and change of the grid load directly affect the power generation plan and power output of the photovoltaic site; environmental factors include natural factors such as light intensity, temperature, humidity, and wind speed, and these factors directly affect the power generation efficiency and performance of the photovoltaic modules; while the layout topology configuration of the photovoltaic site refers to the layout and connection methods of key components such as photovoltaic modules, inverters, and energy storage devices within the photovoltaic site. This configuration determines the path of power transmission and conversion, and has an important impact on the overall performance and stability of the system. The foreground conditions refer to those factors that are relatively flexible, adjustable, and used to optimize the operation performance of the photovoltaic site. When retrieving control samples through network connection, these factors are set as the optimization objectives for retrieval, and are used to find the optimal control parameters that meet specific requirements. When retrieving control samples through network connection, the system will search and match in the historical database according to the set background conditions and foreground conditions, and they jointly determine the direction and objective of the retrieval to find historical data or cases that meet the requirements. Then, the key characteristic values (such as power factor characteristic value, control stability characteristic value, and dynamic response speed characteristic value) in these historical data or cases are extracted for evaluation and comparison to determine the optimal control parameters. Among them, the control stability characteristic value is a particularly important indicator, which is characterized by the minimum eigenvalue of the system. In a photovoltaic grid-connected system, the minimum eigenvalue reflects the stability of the system when facing disturbances. When the minimum eigenvalue approaches 0, the system may be in an unstable state, so special attention needs to be paid to this indicator. With these characteristic values, the fitness evaluation function is used to process them to obtain the fitness indicator of the updated parameters, and this indicator can reflect the comprehensive performance of the parameters under the current conditions. Next, the fitness indicator of the updated parameters is compared with the fitness indicator of the parameters in the initial population.If the fitness identifier of the updated parameter is less than or equal to the minimum fitness identifier in the initial population, it indicates that the current iterative update has not brought about performance improvement, and it is necessary to continue the loop iteration based on the initial population. If the fitness identifier of the updated parameter is greater than the minimum fitness identifier in the initial population, it indicates that the current iterative update is effective. Then, replace the parameter with the minimum fitness identifier in the initial population with the updated device control parameter to obtain a new population, and continue the loop iteration. This iterative process will continue until the preset number of iterations is met. Finally, select the device control parameter with the largest fitness identifier from the population as the optimal control parameter under the current conditions for output. In summary, the entire process is essentially achieved through network optimization by an optimization algorithm. By retrieving and analyzing big data, the device control parameters suitable for the current conditions can be found more accurately, improving the stability and accuracy of control and avoiding the contingency of historical decisions. This strategy can not only improve the operation efficiency of photovoltaic sites but also reduce the operation and maintenance costs, providing strong support for the intelligent management of the photovoltaic industry.

[0032] In a preferred embodiment, taking the multiple groups of initial device control parameters as the initial population and performing population iteration according to the multiple fitness identifiers to obtain updated device control parameters includes: setting search steps for the multiple groups of initial device control parameters according to the multiple fitness identifiers to obtain multiple search steps, where the principle of search step setting is: the larger the fitness, the smaller the search step; distributing the multiple groups of initial device control parameters to the particle distribution space to obtain multiple control particle distribution coordinates, and setting search directions for the multiple control particle distribution coordinates according to the multiple fitness identifiers to obtain multiple search directions, where the principle of search direction setting is: taking the control particle distribution coordinate with the maximum fitness value within the preset distance sphere neighborhood centered on any control particle distribution coordinate as its search direction, and when the center of the sphere neighborhood is the maximum fitness value, taking a random position within the sphere neighborhood as its search direction; updating the multiple groups of initial device control parameters according to the multiple search directions and the multiple search steps to obtain the updated device control parameters.

[0033] Furthermore, in the process of optimizing the control parameters of photovoltaic site equipment, an iterative search method based on fitness is adopted. First, multiple groups of initial equipment control parameters are used as the initial population, and these parameters represent different operation strategies and configurations. Then, the fitness identifiers of these initial parameters are calculated according to their performance in actual operation, and these identifiers reflect the influence degree of the parameter combination on the system performance. Next, the search step size is set for each group of initial equipment control parameters according to the fitness identifiers. The principle of setting the search step size here is that the higher the fitness of a solution, the greater the probability of having a better solution nearby. Therefore, the search step size is set smaller to conduct a detailed search near the potential better solution. On the contrary, for the solution with a smaller fitness, although the probability of having a better solution nearby is relatively small, it still needs to be given a certain search opportunity, but the search step size will be relatively larger to explore in a wider range. Subsequently, multiple groups of initial equipment control parameters are distributed into a particle distribution space, and each parameter combination corresponds to a control particle distribution coordinate. In this space, the search direction is set for each control particle according to the fitness identifier. The principle of setting the search direction is that if there is a control particle distribution coordinate with a higher fitness within the preset distance sphere neighborhood centered on any control particle distribution coordinate, then this coordinate is used as the search direction of the current control particle. If the center of the sphere neighborhood itself is the maximum fitness value, then in order to avoid falling into the local optimal solution, a position within the sphere neighborhood can be randomly selected as its search direction. Finally, multiple groups of initial equipment control parameters are updated according to the set multiple search directions and search step sizes, thereby obtaining the updated equipment control parameters. The fitness of these updated parameter combinations usually improves because they are obtained by conducting a detailed search near the potential better solution or exploring in a wider range. In summary, the entire process conducts iterative search based on fitness, gradually approaching the optimal solution by dynamically adjusting the search step size and search direction, thereby optimizing the control parameters of photovoltaic site equipment.

[0034] In a preferred embodiment, the process of obtaining the updated parameter fitness identifier by processing the power factor eigenvalue, the control stability eigenvalue, and the dynamic response speed eigenvalue through the fitness evaluation function includes: when there are abnormal control records exceeding a preset number in the historical cross-sectional data of the power factor eigenvalue, the control stability eigenvalue, and the dynamic response speed eigenvalue, the updated parameter fitness identifier is set to 0; otherwise, the power factor eigenvalue, the control stability eigenvalue, and the dynamic response speed eigenvalue are processed through the fitness evaluation function to obtain the updated parameter fitness identifier; the fitness evaluation function is: ; where represents the control stability eigenvalue, represents the dynamic response speed eigenvalue, Characterize the power factor eigenvalue, , and characterize the weight parameters, characterize the expected value of control stability, characterize the expected value of dynamic response speed, characterize the expected value of power factor, e represents the natural constant, and fit() represents the fitness evaluation function, which is used to evaluate the comprehensive eigenvalue of the stability eigenvalue, dynamic response speed eigenvalue and power factor eigenvalue.

[0035] Specifically, in the process of optimizing the control parameters of photovoltaic site equipment, a fitness evaluation function is introduced to quantify the performance of the updated parameter combination in actual operation. This function comprehensively considers three key indicators: the power factor eigenvalue, the control stability eigenvalue and the dynamic response speed eigenvalue. First, it is possible to check whether there are abnormal control records of these three eigenvalues in the historical cross-sectional data. Abnormal control records may indicate that the parameter combination cannot work stably or has poor performance under certain conditions. If any one of the power factor eigenvalue, control stability eigenvalue and dynamic response speed eigenvalue appears abnormally in the historical data more than the preset number of times, then this set of updated parameters can be considered unqualified, and its fitness identifier is directly set to 0. Such a setting is to ensure that only parameter combinations that perform stably in actual operation can be further considered. If none of these three eigenvalues have abnormal control records in the historical data, then the fitness evaluation function will be used to calculate the fitness identifier of the updated parameters. This function is in the form of a weighted sum, where each eigenvalue is scored according to the difference between its actual value and the expected value, and is weighted by the corresponding weight parameter. The specific formula is: . Among them, in the fitness evaluation function , and represent the actual values of the control stability eigenvalue, dynamic response speed eigenvalue and power factor eigenvalue respectively, while , and represent the expected values of these three eigenvalues respectively. These expected values are set according to the operation requirements and performance objectives of the photovoltaic site. In addition, , and characterize the weight parameters, which are used to adjust the relative importance of different eigenvalues in the fitness evaluation. e represents the natural constant. Finally, the fitness evaluation function will return a fitness identifier between 0 and the maximum value, which reflects the comprehensive performance of the updated parameter combination in terms of power factor, control stability and dynamic response speed. By comparing the fitness identifiers of different parameter combinations, the optimal parameter combination can be selected as the final optimization result.

[0036] For a more intuitive understanding, the following is an example of the fitness identification process calculated by the fitness evaluation function. The specific representation symbols in the example are different from those in the fitness formula, as shown in the following table:

[0037]

[0038] In the above table, the columns of "Power factor eigenvalue (X)", "Control stability eigenvalue (Y)", and "Dynamic response speed eigenvalue (Z)" respectively represent the actual performance of different parameter combinations in the historical cross-sectional data. The column of "Abnormal control record count" indicates whether there is an abnormal control record for this parameter combination in the historical data. If the number of abnormal records exceeds the preset number (for example, set to 1), the fitness identification is set to 0. The column of "Fitness evaluation function result (F)" represents the fitness score calculated according to the fitness evaluation function. The column of "Fitness identification" is obtained by comprehensively considering the fitness evaluation function result and the number of abnormal control records. If the number of abnormal records exceeds the preset number, the fitness identification is "low" (or 0). Otherwise, according to the result of the fitness evaluation function, the fitness identification is divided into "high", "medium", or "low" (or a specific numerical range).

[0039] In summary, by comprehensively considering the optimization objectives, historical data, system requirements, and potential improvement space of the photovoltaic site, setting reasonable expected values and weight parameters, and using the fitness evaluation function to quantitatively evaluate different parameter combinations, the technical effect of accurately guiding the optimization of the device control parameters of the photovoltaic site is achieved, effectively improving the power generation efficiency, stability, and dynamic response speed of the system.

[0040] In a preferred embodiment, obtaining the device control parameters for state mapping to obtain the baseline state of device operation includes: the device control parameters are processed by a state mapping network corresponding to the photovoltaic site to output the baseline state of device operation; the state mapping network is trained based on machine learning with multiple groups of data, where any one of the multiple groups of data includes: device control parameter record data and label data identifying the baseline state of device operation.

[0041] Exemplarily, in the process of optimizing the control parameters of photovoltaic site equipment, in order to obtain the baseline operating state of the equipment, a means based on a state mapping network is adopted. The core of this operation lies in quickly and accurately processing the equipment control parameters through the state mapping network, so as to output the baseline operating state of the equipment. Specifically, the state mapping network is trained through machine learning techniques. To train this network, multiple sets of data are first collected. Among them, each set of data contains the recorded data of the equipment control parameters and the label data indicating the baseline operating state of the equipment. The collection sources of these data are mainly the real-time monitoring systems of photovoltaic sites, which record the control parameters and corresponding baseline states of the equipment under different operating conditions. After sufficient data is collected, a suitable machine learning algorithm and model are selected to construct the state mapping network. Commonly used machine learning algorithms include neural networks, support vector machines, etc. In this scenario, a deep learning model such as a convolutional neural network (CNN) or a recurrent neural network (RNN) can be selected because these models perform well in processing complex data and pattern recognition. Next, the collected data is used to train the state mapping network. The training process includes inputting the data into the network, calculating the output through forward propagation, and then comparing it with the actual label data to calculate the error. Then, the backpropagation algorithm is used to adjust the weights and biases of the network to reduce the error. This process is repeated multiple times until the performance of the network reaches a stable level, that is, it can accurately output the baseline operating state of the equipment according to the equipment control parameters. The trained state mapping network can quickly process new equipment control parameters and output the corresponding baseline operating state of the equipment. Since the network has learned the patterns and rules in a large amount of data, it can perform well on new data and has high mapping efficiency and accuracy. In summary, the state mapping network can quickly process a large number of equipment control parameters and output the baseline operating state of the equipment in real time, greatly improving the mapping efficiency; the state mapping network obtained through machine learning training can accurately identify the mapping relationship between the equipment control parameters and the baseline operating state, thus improving the mapping accuracy; in addition, after obtaining the accurate baseline operating state of the equipment, the equipment control parameters can be further optimized to improve the power generation efficiency, stability and dynamic response speed of the photovoltaic site.

[0042] In a preferred embodiment, the multiple sets of data collection steps include: obtaining the recorded data of the equipment control parameters of the photovoltaic site; retrieving several groups of historical operating state monitoring data of normal operation based on the recorded data of the equipment control parameters; performing an interval analysis of the same-attribute state data sets on the several groups of historical operating state monitoring data to obtain the label data indicating the baseline operating state of the equipment.

[0043] Optionally, to extract valuable information from the historical operation data of the photovoltaic site and provide a solid foundation for subsequent machine learning tasks, the recorded data of device control parameters is first collected from the monitoring system of the photovoltaic site. These data usually include key parameters such as voltage, current, power factor, temperature, etc., which reflect the status of the device under different operating conditions. The collection of these data is real-time to ensure that the latest dynamics of the device operation can be captured. Next, based on the recorded data of device control parameters collected, the historical database of the photovoltaic site is used to retrieve the normal operation status monitoring data corresponding to these parameters. These historical data record the normal operation status of the device over a period of time in the past (such as the past few months or years), including the performance of the device under various working conditions. By selecting the operation status data marked as "normal" or "stable", the accuracy and reliability of subsequent analysis are ensured. After obtaining several groups of normal operation status monitoring data, the centralized interval analysis of the same-attribute status data is carried out. The purpose of this step is to identify the value range or distribution characteristics of each control parameter of the device in the normal state. The distribution of these data is analyzed through statistical methods (such as mean, standard deviation, maximum value, minimum value, etc.), and a reasonable interval range is determined, which can cover most of the data points in the normal state, that is, the centralized interval of the data. Then, based on the results of the centralized interval analysis of the same-attribute status data, the label data indicating the baseline status of the device operation can be generated. For each group of historical operation status monitoring data, check whether its control parameters fall within the previously determined normal value range. If all or most of the parameters fall within this range, the group of data is marked as "normal" or "baseline status". On the contrary, if some parameters deviate from the normal range, the group of data may be marked as "abnormal" or "non-baseline status". To sum up, the multiple sets of data sets with device operation baseline status labels extracted from the historical operation data of the photovoltaic site not only provide valuable label data for the training of the status mapping network, but also provide an important reference basis for the subsequent optimization of device control parameters. It should be emphasized that the entire data collection process requires strict quality control to ensure the accuracy, integrity and consistency of the data. In addition, as the operating conditions of the photovoltaic site continue to change and new data are continuously generated, it is also necessary to regularly update and optimize the data collection process to adapt to new requirements and challenges.

[0044] In a preferred embodiment, constructing a device state fluctuation curve for co-modal high-frequency abnormal period indexing to obtain a device state monitoring period, including: networking to index the device state fluctuation record curve before a trigger fault when the number of trigger faults in a photovoltaic power station exceeds a trigger fault number threshold, wherein the length of the first time zone of the device state fluctuation record curve is greater than the length of the second time zone of the device state fluctuation curve; performing differential feature extraction on the device state fluctuation record curve to obtain a first differential feature sequence; performing differential feature extraction on the device state fluctuation curve to obtain a second differential feature sequence; according to the second differential feature sequence, intercepting an equi-length sequence from the first differential feature sequence for Euclidean distance comparison to obtain a Euclidean distance evaluation value; when the Euclidean distance evaluation value is less than or equal to a Euclidean distance threshold, statistically storing the interval duration between the end time of the intercepted sequence and the trigger fault time; when the storage quantity of the interval duration is greater than or equal to a preset statistical quantity, taking the minimum value of the non-outlier points of all the interval durations as the device state monitoring period.

[0045] Specifically, devices with the number of triggered faults exceeding a preset threshold are screened out from the historical database of the photovoltaic power station, and the state fluctuation record curves of these devices before each triggered fault are obtained. These curves record the changes in key parameters such as voltage, current, power, etc. of the device during a period of time (such as several hours, days or weeks) before the fault occurs. It should be noted that in order to capture the state changes on different time scales, two time zone lengths are set here: the first time zone length is longer and is used to capture long-term trends; the second time zone length is shorter and is used to capture short-term fluctuations. Next, differential feature extraction is performed on each device state fluctuation record curve. The differential feature refers to the change amount between parameter values at adjacent time points, which can reflect the change rate and trend of the device state. Differential feature extraction is performed on the curves with the first time zone length and the second time zone length respectively to obtain two groups of differential feature sequences: the first differential feature sequence and the second differential feature sequence. Then, using the second differential feature sequence as a reference, a sequence of the same length as it is intercepted from the first differential feature sequence for Euclidean distance comparison. The Euclidean distance is an index to measure the similarity between two vectors, and the smaller the distance, the more similar the two vectors are. Through comparison, long-term trend segments similar to short-term fluctuations can be found and the similarity between them can be evaluated. When the Euclidean distance evaluation value is less than or equal to the preset Euclidean distance threshold, it is considered that this long-term trend and short-term fluctuation have the same modal characteristics and may indicate an impending fault. At this time, the time interval between the end time of the intercepted sequence and the triggered fault time is statistically counted and stored, and this time interval reflects the time interval from the abnormal state of the device to the actual occurrence of the fault. Finally, when the number of stored time intervals reaches the preset statistical number, these time intervals are analyzed. In order to exclude the influence of outliers (i.e., abnormal values), the minimum value of the non-outlier of all time intervals is taken as the device state monitoring period. This period represents the shortest time interval in history from the abnormal state of the device to the occurrence of the fault. Finally, taking the shortest interval of historical trigger accidents as the monitoring duration can greatly increase the probability of timely monitoring the abnormal state of the device. Because within this period, the change of the device state is the most intense and it is also most likely to expose potential fault hazards. By regularly monitoring the device state and taking corresponding preventive measures, the incidence of device faults can be effectively reduced, and the operation efficiency and safety of the photovoltaic power station can be improved. In summary, the method of constructing a device state fluctuation curve for indexing the same modal high-frequency abnormal cycle provides a scientific and effective device state monitoring means. It can not only timely detect the abnormal state of the device, but also guide the formulation of reasonable maintenance plans and preventive measures, providing a strong guarantee for the long-term stable operation of the photovoltaic power station.

[0046] Embodiment 2:

[0047] Such as Figure 2As shown, based on the same inventive concept as the integrated control method for distributed photovoltaics provided in Embodiment 1, an embodiment of the present invention further provides an integrated control device for distributed photovoltaics, and the device includes:

[0048] A parameter acquisition module 11, configured to obtain device control parameters for state mapping and obtain the baseline state of device operation when the photovoltaic site does not trigger the islanding effect.

[0049] A state discrimination module 12, configured to receive device state information at a default period. If the device state information is consistent with the baseline state of device operation, construct a device state fluctuation curve for indexing the high-frequency abnormal period of the same mode, and obtain an update of the default period of the device state monitoring period.

[0050] An off-grid control module 13, configured to perform off-grid control of the photovoltaic site through a photovoltaic circuit breaker when the photovoltaic site triggers the islanding effect, or when the device state information is inconsistent with the baseline state of device operation.

[0051] Furthermore, the parameter acquisition module 11 is further configured to perform the following steps:

[0052] When the photovoltaic site does not trigger the islanding effect, when the first fluctuation value between the monitored grid load and the pre-stored grid load is greater than or equal to the first fluctuation threshold, and / or the second fluctuation value between the monitored environmental factors and the pre-stored environmental factors is greater than or equal to the second fluctuation threshold, input the monitored grid load and the monitored environmental factors into the historical control database to retrieve device control parameters; use the monitored grid load to update the pre-stored grid load, use the monitored environmental factors to update the pre-stored environmental factors, and at the same time send the device control parameters to the photovoltaic site for control initialization.

[0053] Furthermore, the parameter acquisition module 11 is further configured to perform the following steps:

[0054] Input the grid load and the environmental factors into the historical control database, and retrieve multiple groups of initial device control parameters of the photovoltaic site. Among them, the multiple groups of initial device control parameters have multiple fitness identifiers. The greater the fitness, the better the control parameters. Use the multiple groups of initial device control parameters as the initial population, and perform population iteration according to the multiple fitness identifiers to obtain updated device control parameters. Input the grid load, the environmental factors, and the background conditions of the photovoltaic site layout topology configuration, and set the updated device control parameters as the foreground conditions. Retrieve the power factor eigenvalue, control stability eigenvalue, and dynamic response speed eigenvalue of the control samples that meet the background conditions and the foreground conditions through networking. Process the power factor eigenvalue, the control stability eigenvalue, and the dynamic response speed eigenvalue through a fitness evaluation function to obtain an updated parameter fitness identifier. When the updated parameter fitness identifier is less than or equal to the minimum value of the multiple fitness identifiers, perform cyclic iteration based on the multiple groups of initial device control parameters. When the updated parameter fitness identifier is greater than the minimum value of the multiple fitness identifiers, use the updated device control parameters to replace the initial device control parameters with the minimum fitness identifier among the multiple groups of initial device control parameters, and obtain an updated population for cyclic iteration. When the number of cyclic iterations meets the preset number, output the device control parameters with the maximum fitness identifier.

[0055] Furthermore, the parameter acquisition module 11 is further configured to perform the following steps:

[0056] According to the multiple fitness identifiers, set search step sizes for the multiple groups of initial device control parameters to obtain multiple search step sizes. The principle of search step size setting is: the greater the fitness, the smaller the search step size. Distribute the multiple groups of initial device control parameters to the particle distribution space to obtain multiple control particle distribution coordinates. Set search directions for the multiple control particle distribution coordinates according to the multiple fitness identifiers to obtain multiple search directions. The principle of search direction setting is: use the control particle distribution coordinate with the maximum fitness within the preset distance sphere neighborhood centered on any control particle distribution coordinate as its search direction. When the center of the sphere neighborhood is the maximum fitness, use a random position within the sphere neighborhood as its search direction. Update the multiple groups of initial device control parameters according to the multiple search directions and the multiple search step sizes to obtain the updated device control parameters.

[0057] Furthermore, the parameter acquisition module 11 is further configured to perform the following steps:

[0058] When the power factor eigenvalue, the control stability eigenvalue, and the dynamic response speed eigenvalue have abnormal control records exceeding a preset number in the historical cross-sectional data, the update parameter fitness flag is set to 0; otherwise, the power factor eigenvalue, the control stability eigenvalue, and the dynamic response speed eigenvalue are processed according to the fitness evaluation function to obtain the update parameter fitness flag; the fitness evaluation function is: ; where represents the control stability eigenvalue, represents the dynamic response speed eigenvalue, represents the power factor eigenvalue, 、 and represent the weight parameters, represents the control stability expectation value, represents the dynamic response speed expectation value, represents the power factor expectation value, e represents the natural constant, and fit() represents the fitness evaluation function, which is used to evaluate the comprehensive eigenvalue of the stability eigenvalue, the dynamic response speed eigenvalue, and the power factor eigenvalue.

[0059] Furthermore, the parameter acquisition module 11 is further configured to perform the following steps:

[0060] The device control parameters are processed by the state mapping network corresponding to the photovoltaic site to output the baseline state of the device operation; the state mapping network is obtained by machine learning training based on multiple sets of data, where any set of the multiple sets of data includes: device control parameter record data and label data identifying the baseline state of the device operation.

[0061] Furthermore, the parameter acquisition module 11 is further configured to perform the following steps:

[0062] Obtain the device control parameter record data of the photovoltaic site; retrieve several groups of historical operation state monitoring data of normal operation based on the device control parameter record data; perform interval analysis on the same attribute state data in the several groups of historical operation state monitoring data to obtain the label data identifying the baseline state of the device operation.

[0063] Furthermore, the state discrimination module 12 is further configured to perform the following steps:

[0064] The device state fluctuation record curve before a trigger fault where the number of trigger faults of an Internet-connected indexed photovoltaic power station exceeds the trigger fault number threshold, wherein the length of the first time zone of the device state fluctuation record curve is greater than the length of the second time zone of the device state fluctuation curve; perform differential feature extraction on the device state fluctuation record curve to obtain a first differential feature sequence; perform differential feature extraction on the device state fluctuation curve to obtain a second differential feature sequence; according to the second differential feature sequence, intercept an equal-length sequence from the first differential feature sequence for Euclidean distance comparison to obtain a Euclidean distance evaluation value; when the Euclidean distance evaluation value is less than or equal to the Euclidean distance threshold, statistically store the interval duration between the end time of the intercepted sequence and the trigger fault time; when the storage quantity of the interval duration is greater than or equal to the preset statistical quantity, take the minimum value of the non-outlier of all the interval durations and set it as the device state monitoring period.

[0065] Through the foregoing detailed description of a fusion control method for a distributed photovoltaic, those skilled in the art can clearly know a fusion control device for a distributed photovoltaic in this embodiment. For the device disclosed in the embodiment, since it corresponds to the method disclosed in the embodiment, the description is relatively simple. For the relevant parts, refer to the description in the method part.

[0066] The above description of the disclosed embodiments enables those skilled in the art to implement or use the present application. Various modifications to these embodiments will be obvious to those skilled in the art. The general principles defined herein can be implemented in other embodiments without departing from the spirit or scope of the present application. Therefore, the present application will not be limited to the embodiments shown herein, but rather to the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A distributed photovoltaic fusion control method, characterized in that: Applied to the master station, the master station communicates with the photovoltaic site, including: When the islanding effect is not triggered at the photovoltaic site, the equipment control parameters are obtained for state mapping to obtain the equipment operation baseline state; Receive device status information according to a default cycle, and if the device status information is consistent with the device operating baseline state, construct a device status fluctuation curve to perform a same-mode high-frequency abnormal cycle index, and obtain a device status monitoring cycle to update a default cycle; When an islanding effect is triggered at a photovoltaic site, or the device status information is inconsistent with the device operation baseline status, off-grid control of the photovoltaic site is performed through a photovoltaic circuit breaker.

2. The method according to claim 1, characterized in that Get device control parameters, including: When the photovoltaic site does not trigger the islanding effect, when the first fluctuation value between the monitored grid load and the pre-stored grid load is greater than or equal to the first fluctuation threshold, or / and the second fluctuation value between the monitored environmental factor and the pre-stored environmental factor is greater than or equal to the second fluctuation threshold, the monitored grid load and the monitored environmental factor are input into the historical control database to retrieve the device control parameter; The pre-stored grid load is updated using the monitored grid load, the pre-stored environmental factor is updated using the monitored environmental factor, and the device control parameter is sent to the photovoltaic site for control initialization.

3. The method according to claim 2, characterized in that Input grid load and environmental factors into the historical control database to retrieve equipment control parameters, including: Inputting the grid load and the environmental factors into the historical control database, retrieving multiple sets of initial device control parameters of the photovoltaic site, wherein the multiple sets of initial device control parameters have multiple fitness identifiers, and the greater the fitness, the better the control parameters; Taking the multiple groups of initial device control parameters as the initial population, performing population iteration according to the multiple fitness identifiers to obtain updated device control parameters; The grid load, the environmental factors and the photovoltaic site layout topology are configured as background conditions, the update device control parameters are set as foreground conditions, and the power factor characteristic value, control stability characteristic value and dynamic response speed characteristic value of the control samples that meet the background conditions and the foreground conditions are retrieved online; Processing the power factor characteristic value, the control stability characteristic value and the dynamic response speed characteristic value through a fitness evaluation function to obtain an update parameter fitness identifier; When the update parameter fitness identifier is less than or equal to the minimum value of the multiple fitness identifiers, performing loop iteration based on the multiple groups of initial device control parameters; When the fitness identifier of the update parameter is greater than the minimum value of the multiple fitness identifiers, the initial device control parameter of the minimum value of the fitness identifiers of the multiple groups of initial device control parameters is replaced by the update device control parameter to obtain an update population for cyclic iteration; When the number of loop iterations meets the preset number, the device control parameter with the maximum fitness indicator is output.

4. The method according to claim 3, characterized in that Taking the multiple groups of initial device control parameters as the initial population, performing population iteration according to the multiple fitness identifiers to obtain updated device control parameters, including: According to the multiple fitness identifiers, search step lengths are set for the multiple groups of initial device control parameters to obtain multiple search step lengths, wherein the search step length setting principle is: the greater the fitness, the smaller the search step length; The multiple groups of initial device control parameters are distributed to the particle distribution space to obtain multiple control particle distribution coordinates, and the search directions are set for the multiple control particle distribution coordinates according to the multiple fitness identifiers to obtain multiple search directions, wherein the search direction setting principle is: the control particle distribution coordinate with the maximum fitness value within a spherical neighborhood with a preset distance from the center of any control particle distribution coordinate is used as its search direction, and when the center of the spherical neighborhood is the maximum fitness value, a random position in the spherical neighborhood is used as its search direction; The multiple groups of initial device control parameters are updated according to the multiple search directions and the multiple search step sizes to obtain the updated device control parameters.

5. The method according to claim 3, characterized in that Processing the power factor characteristic value, the control stability characteristic value, and the dynamic response speed characteristic value through a fitness evaluation function to obtain an update parameter fitness identifier includes: When the power factor characteristic value, the control stability characteristic value and the dynamic response speed characteristic value have abnormal control records exceeding a preset number of times in the historical cross-sectional data, the update parameter fitness flag is set to 0; Otherwise, the power factor characteristic value, the control stability characteristic value and the dynamic response speed characteristic value are processed according to the fitness evaluation function to obtain an update parameter fitness identifier; The fitness evaluation function is: in, Characterize the control stability eigenvalue, Characterize the dynamic response speed characteristic value, Characterize the power factor characteristic value, , and Characterize the weight parameters, Characterizes the expected value of control stability, Characterizes the expected value of dynamic response speed, represents the expected value of the power factor, e represents the natural constant, and fit() represents the fitness evaluation function, which is used to evaluate the comprehensive characteristic value of the stability characteristic value, the dynamic response speed characteristic value and the power factor characteristic value.

6. The method according to claim 1, characterized in that Obtain equipment control parameters for status mapping and obtain equipment operation baseline status, including: The equipment control parameters are processed by a state mapping network corresponding to the photovoltaic site, and the equipment operation baseline state is output; The state mapping network is obtained through multiple sets of data based on machine learning training, wherein any set of the multiple sets of data includes: equipment control parameter recording data and label data identifying the baseline state of equipment operation.

7. The method according to claim 6, characterized in that The multiple sets of data collection steps include: Obtain equipment control parameter record data of photovoltaic sites; Retrieving several groups of historical operation status monitoring data of normal operation based on the equipment control parameter record data; The plurality of groups of historical operating status monitoring data are subjected to interval analysis of the same attribute status data to obtain label data identifying the operating baseline status of the equipment.

8. The method according to claim 1, characterized in that Construct the equipment status fluctuation curve to index the same-mode high-frequency abnormal period and obtain the equipment status monitoring period, including: The equipment state fluctuation record curve before the triggering fault of the networking index photovoltaic power station exceeds the triggering fault number threshold, wherein the first time zone length of the equipment state fluctuation record curve is greater than the second time zone length of the equipment state fluctuation curve; Performing differential feature extraction on the equipment state fluctuation record curve to obtain a first differential feature sequence; Performing differential feature extraction on the device state fluctuation curve to obtain a second differential feature sequence; According to the second differential feature sequence, a sequence of equal length is intercepted from the first differential feature sequence to perform Euclidean distance comparison to obtain a Euclidean distance evaluation value; When the Euclidean distance evaluation value is less than or equal to the Euclidean distance threshold, the interval between the end point of the interception sequence and the fault triggering time is statistically collected and stored; When the stored number of interval durations is greater than or equal to the preset statistical number, the minimum value of non-outlier points of all interval durations is taken as the device status monitoring period.

9. A distributed photovoltaic fusion control device, characterized in that: Applied to a master station, the master station is connected to a photovoltaic site for communication, and is used to implement a distributed photovoltaic integration control method according to any one of claims 1 to 8, the device comprising: The parameter acquisition module is used to obtain the equipment control parameters for state mapping and obtain the equipment operation baseline state when the photovoltaic site does not trigger the islanding effect; A state identification module is used to receive device state information according to a default period. If the device state information is consistent with the device operation baseline state, a device state fluctuation curve is constructed to perform a same-mode high-frequency abnormal cycle index, and a default period for updating the device state monitoring cycle is obtained; The off-grid control module is used to perform off-grid control of the photovoltaic site through a photovoltaic circuit breaker when an islanding effect is triggered at the photovoltaic site or the device status information is inconsistent with the device operation baseline status.

Images