An artificial intelligence-based management system for photovoltaic power generation

Abstract

The application relates to the technical field of photovoltaic power generation operation management, and discloses a management system for photovoltaic power generation based on artificial intelligence, which comprises a data acquisition unit, an operation state modeling unit, an operation influence correlation information updating unit, an operation management strategy generating unit and a closed-loop management unit. The operation state model is established by collecting the operation data of photovoltaic components, inverters, energy storage devices, grid-connected devices and environment monitoring devices, and the operation influence correlation information is dynamically updated according to operation feedback, so that the operation state model continuously reflects the changes of the operation influence correlation information between various operation objects; the operation management strategy generating unit generates an operation management strategy according to the updated operation state model, and realizes that operation feedback continuously participates in the generation of the management strategy of the next operation cycle through closed-loop management, thereby improving the dynamic adaptability, operation stability and management efficiency of photovoltaic power generation system operation management.

Description

Technical Field

[0001] This invention relates to the field of photovoltaic power generation operation and management technology, and in particular to a management system for photovoltaic power generation based on artificial intelligence. Background Technology

[0002] As the installed capacity of photovoltaic power generation continues to increase, photovoltaic power generation systems have gradually evolved from single power generation equipment into comprehensive operating systems composed of photovoltaic modules, inverters, energy storage equipment, grid connection equipment, and environmental monitoring equipment. In order to improve power generation efficiency and operational stability, it is usually necessary to conduct unified monitoring of each operating object and to schedule and manage the operation of the photovoltaic power generation system based on the monitoring results.

[0003] The existing technology still has the following shortcomings: 1. Most existing photovoltaic power generation management systems use fixed relationships or static operating rules for management, which makes it difficult for the operation management strategy to reflect the dynamic changes in the relationship between various operating objects in a timely manner. When the operating environment changes, the operation management strategy is prone to be inconsistent with the actual operating status of the system, thereby affecting the overall operating efficiency of the photovoltaic power generation system.

[0004] 2. After completing one operation scheduling, the existing photovoltaic power generation management system usually only generates the updated management strategy for the next cycle based on the new real-time operation data. It cannot continuously feed back the changes in the impact of the previous operation cycle's execution results on the relationship between the operating objects to the subsequent operation management process. This makes the operation management process lack continuous dynamic adjustment capabilities, which can easily reduce the adaptability of the management strategy and the stability of operation under complex operating conditions. Summary of the Invention

[0005] To address the shortcomings of existing technologies, this invention provides an artificial intelligence-based management system for photovoltaic power generation. This system solves the problems of existing photovoltaic power generation management systems, which rely on fixed operational relationships for operation management, making it difficult to reflect the dynamic changes in the relationships between operational objects, and lacking a mechanism for continuous updating of operational relationship information, resulting in insufficient adaptability of operation management strategies.

[0006] To achieve the above objectives, the present invention adopts the following technical solution: An artificial intelligence-based photovoltaic power generation management system includes: a data acquisition unit for collecting operational data from photovoltaic power generation equipment, energy storage equipment, grid-connected equipment, and environmental monitoring equipment in the photovoltaic power generation system; The operation status modeling unit is used to establish an operation status model of the photovoltaic power generation system based on the operation data. The operation status model is used to uniformly describe the operation status of each operation object. The operation impact correlation information update unit is used to establish operation impact correlation information between each operation object based on the operation status model, and dynamically update the operation impact correlation information based on operation feedback to reflect the changes in operation impact correlation information between each operation object. The management strategy generation unit is used to generate corresponding operation management strategies based on the operation status model and the updated operation impact correlation information. The closed-loop management unit is used to obtain the operation feedback after the operation management strategy is executed, and drive the update of the operation impact correlation information based on the operation feedback, so that the updated operation impact correlation information participates in the generation of the management strategy for the next operation cycle, thereby forming a closed-loop management of the photovoltaic power generation system driven by the operation status model.

[0007] Furthermore, the operating data includes photovoltaic module output parameters, inverter operating parameters, energy storage device operating parameters, grid-connected operating parameters, and environmental monitoring parameters, wherein the environmental monitoring parameters include at least one or more of solar irradiance, ambient temperature, module temperature, and wind speed.

[0008] Furthermore, the operating state model includes equipment operating state, energy transfer state, and operating sequence state, which are used to describe the current operating state of each operating object and its corresponding operating attributes.

[0009] Furthermore, the operational impact correlation information includes at least one or more of the following: power transfer relationship, energy exchange relationship, control response relationship, and scheduling dependency relationship.

[0010] Furthermore, the operation impact association information update unit dynamically adjusts the operation impact association information based on operation feedback. The dynamic adjustment includes at least one or more of the following: association weight update, association direction update, association validity update, and association priority update.

[0011] Furthermore, the management strategy generation unit generates at least one of the following based on the operating status model and operating impact correlation information: photovoltaic array operating mode, energy storage charging and discharging strategy, inverter control strategy, grid connection control strategy, and operation and maintenance strategy.

[0012] Furthermore, the closed-loop management unit continuously acquires operational feedback and forms a feedback data stream.

[0013] Furthermore, the operational feedback includes one or more of the following: power generation change information, energy storage status change information, grid connection status change information, equipment response information, and environmental change information.

[0014] Furthermore, after the operation impact association information is updated, it re-participates in the operation management strategy generation to ensure that the newly generated operation management strategy is consistent with the current operation impact association information.

[0015] Furthermore, the operational impact correlation information adopts a data organization structure of correlation matrix.

[0016] Compared with the prior art, the beneficial effects of the present invention are as follows: 1. This invention establishes a unified operating state model to uniformly describe each operating object of the photovoltaic power generation system, and continuously maintains the operating impact correlation information during the operation management process. This enables the operating state model to reflect the dynamic changes in the operating impact correlation information between operating objects, avoiding the problem of insufficient adaptability of operation management strategies caused by the use of fixed operating relationships in the existing technology.

[0017] 2. This invention dynamically updates operational impact correlation information based on operational feedback, enabling the operational feedback generated in the previous operational cycle to continuously affect subsequent operational management processes, thereby achieving continuous optimization of the operational state model, improving the consistency between operational management strategies and actual operational states, and enhancing the continuous adjustment capability of operational management under complex operating conditions.

[0018] 3. This invention generates operation management strategies based on the updated operation status model and applies the operation feedback to the operation status model again, forming a closed-loop operation management mechanism driven by the operation status model, thereby improving the dynamic adaptability and overall operation stability of the photovoltaic power generation system.

[0019] 4. This invention does not require changes to the original equipment structure of the photovoltaic power generation system. It achieves operation management optimization only through the dynamic maintenance of the operation status model and operation impact information. It is compatible with existing photovoltaic modules, inverters, energy storage equipment and grid-connected equipment, and has good engineering applicability and promotion application value. Attached Figure Description

[0020] Figure 1 This is a schematic diagram of the system structure of a photovoltaic power generation management system based on artificial intelligence proposed in this invention; Figure 2 This is a schematic diagram illustrating the influence relationship of operating objects in a photovoltaic power generation management system based on artificial intelligence proposed in this invention. Figure 3 This is a schematic diagram of the operational impact-related information structure of a photovoltaic power generation management system based on artificial intelligence, as proposed in this invention. Figure 4 This is a schematic diagram illustrating the process of generating an operation management strategy for a photovoltaic power generation management system based on artificial intelligence, as proposed in this invention. Figure 5 This is a schematic diagram of the closed-loop operation management process of a photovoltaic power generation management system based on artificial intelligence proposed in this invention; Figure 6This is a schematic diagram of the overall process of a photovoltaic power generation management system based on artificial intelligence proposed in this invention. Detailed Implementation

[0021] The technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative effort are within the scope of protection of the present invention.

[0022] Example 1, please refer to Figures 1 to 6 This is the first embodiment of the present invention. This embodiment provides a management system for photovoltaic power generation based on artificial intelligence, including a data acquisition unit, an operation status modeling unit, an operation impact correlation information update unit, an operation management strategy generation unit, and a closed-loop management unit.

[0023] The data acquisition unit is connected to photovoltaic modules, inverters, energy storage devices, grid-connected devices, and environmental monitoring devices to obtain real-time operating data of each component of the photovoltaic power generation system.

[0024] The operation status modeling unit is connected to the data acquisition unit and is used to establish a unified operation status model based on the operation data of each operation object. The operation status model is used to describe the current operation status of each operation object and record the operation influence correlation information between each operation object. The operation influence correlation information is used to reflect the degree of influence of a change in one operation object on the operation status of another operation object. The operation status model can be established using machine learning models, deep learning models or other artificial intelligence models.

[0025] The artificial intelligence model uses a multi-layer backpropagation neural network to fit the coupling relationship between various devices. The model training process is as follows: Sample dataset construction: Collect 30 consecutive days of historical operating data as the training set. The sample features include photovoltaic output power, module temperature, irradiance, inverter load, energy storage SOC, grid-connected power, and ambient wind speed; the label is the power fluctuation coupling difference between each pair of devices. Model training rules: Adam gradient descent optimizer is used, learning rate is set to 0.001, iterations are 1000, and mean squared error (MSE) is used as the loss function; after training, the neural network weight file is saved. Online inference process: The data acquisition unit uploads standardized operating parameters to the neural network input layer in real time, and the model inference outputs the original coupling influence coefficients between each pair of devices as the initial values ​​of the correlation matrix; Inference cycle: Synchronized with the system scheduling time window, the model inference output coupling coefficient is executed once every 15 minutes.

[0026] The operation impact correlation information update unit is connected to the operation status modeling unit. It is used to dynamically update the operation impact correlation information in the operation status model based on the operation feedback after the execution of the operation management strategy, so that the operation status model can continuously reflect the changes in the influence relationship between various operating objects during the operation of the photovoltaic power generation system.

[0027] The operation management strategy generation unit is connected to the operation status modeling unit and the operation impact association information update unit, respectively, and is used to generate corresponding operation management strategies based on the operation status model and the updated operation impact association information.

[0028] The closed-loop management unit is connected to the operation management strategy generation unit to obtain operation feedback after the operation management strategy is executed, and returns the operation feedback to the operation impact association information update unit so that the updated operation impact association information participates in the generation of operation management strategy in the next operation cycle, thereby forming a closed-loop operation management process driven by the operation status model.

[0029] Operational status model establishment process: The data acquisition unit continuously collects operational data of each operating object in the photovoltaic power generation system. The operational data includes photovoltaic module output power, voltage, current, module temperature, inverter operating status, energy storage device state of charge, charging and discharging power, grid-connected power, grid status, and environmental parameters such as solar irradiance, ambient temperature, and wind speed.

[0030] The operation status modeling unit establishes an operation status model based on the above operation data. The operation status model includes at least equipment operation status information, energy transfer status information, and operation sequence status information. The equipment operation status information is used to describe the current operation status of each operating object, the energy transfer status information is used to describe the energy flow relationship between operating objects, and the operation sequence status information is used to record the operation change process of the operating object in different operation cycles.

[0031] Based on the established operational status model, the operational status modeling unit further generates operational impact correlation information. This operational impact correlation information describes the operational impact correlation information between various operational objects, including one or more of the following: power transfer impact, energy exchange impact, control response impact, and scheduling dependency impact. This information is stored in the operational status model as an important component, providing an operational basis for the subsequent generation of operational management strategies.

[0032] Example 2, as Figures 3 to 5 As shown, this is the second embodiment of the present invention. Unlike the previous embodiment, this embodiment provides a dynamic update process of operation impact-related information, an operation management strategy generation process, and a closed-loop operation management process of the management system for photovoltaic power generation based on artificial intelligence.

[0033] Dynamic update process of operation impact correlation information: After the operation status model is established, the operation impact correlation information update unit continuously receives operation feedback information from the closed-loop management unit, and dynamically updates the operation impact correlation information in the operation status model based on the operation feedback information.

[0034] Feedback data stream to correlation matrix numerical standardization transformation rules: Input: Structured feedback data stream, with fields including: power of each device before policy execution. Actual power after strategy execution Energy storage SOC change SOC and environmental parameter changes ; Output: Pairwise coupling correction coefficients for each device ; Conversion steps: Calculate the actual fluctuation of a single device ; Match the source object and target object combinations, and calculate the fluctuation coupling difference between them. ; Will Mapped to weight correction increment , which serves as the input for matrix updates.

[0035] The operational feedback information includes one or more of the following: photovoltaic module output change information, inverter operating status change information, energy storage device state of charge change information, grid connection status change information, and environmental parameter change information.

[0036] Operational impact correlation information is used to describe the operational impact correlation information between various operational objects under the current operating conditions. It does not represent the operating parameters of each operational object itself, but rather the degree of impact of a change in the operating state of one operational object on the operating state of another operational object.

[0037] For example, when solar irradiance increases and the output power of photovoltaic modules increases, the operating state of the modules in the operating state model changes, and the degree of operational impact of photovoltaic modules on inverter power regulation, energy storage charging, and grid-connected output changes accordingly. When the state of charge of energy storage devices decreases to a preset range, the degree of operational impact of energy storage devices on grid-connected output capacity and photovoltaic module power absorption capacity also changes.

[0038] Based on the above-mentioned operational feedback information, the operational impact correlation information update unit updates the operational impact correlation information between corresponding operational objects in the operational status model, so that the operational status model continues to be consistent with the current operational conditions without the need to rebuild a new operational status model.

[0039] Weight Iterative Update Formula:

[0040] in: The original weights before the update. For the updated weights, The historical weighting coefficient is retained (fixed value of 0.7). This is the correction increment calculated based on the feedback in this cycle; Matrix update trigger conditions: Operating condition fluctuation trigger threshold: power fluctuation of any device If the rated power is 5% or the single change in energy storage SOC is 3%, matrix update will be initiated immediately. Fixed iteration cycle: When there are no significant fluctuations in operating conditions, a small full matrix update is forcibly performed every 15 minutes of the time sequence window; Association validity update rule: If the coupling coefficient is ≤0.1 for 3 consecutive scheduling cycles, the association is marked as invalid and will not be included in the calculation during the policy generation stage.

[0041] The updated operational impact correlation information is retained in the operational status model and serves as the basis for generating operational management strategies in the next operational cycle. This enables the operational status model to continuously reflect the changes in the impact relationships between operational objects during the operation of the photovoltaic power generation system, thus achieving continuous evolution of the operational status model.

[0042] The data organization structure of the operation impact related information adopts the association matrix.

[0043] ; Line (Source Object): Represents the object that has an impact on the operation; Column (Target Object): Represents the object affected by the execution; Matrix unit: Records the operational impact correlation information between the corresponding source object and the target object, including the direction of impact, the intensity of impact, the effectiveness of impact, and the update time.

[0044] The original coupling coefficients output by the AI ​​inference are normalized to the [0,1] interval, and three levels of strength thresholds are defined: Low impact strength: 0 < coupling coefficient ≤ 0.3; Influence strength: 0.3 < coupling coefficient ≤ 0.7; High impact strength: 0.7 < coupling coefficient ≤ 1; The associated weight is the normalized coupling coefficient, and the matrix cell synchronously stores the precise weight value plus hierarchical labels; The association priority value is equal to the normalized coupling coefficient; the larger the coefficient, the higher the priority.

[0045] Operation management strategy generation process: After the operation impact association information is updated, the operation management strategy generation unit reads the equipment operation status information and operation impact association information in the operation status model, and generates the corresponding operation management strategy based on the operation impact association information between each operation object.

[0046] Operational impact association information includes at least the direction of impact, the intensity of impact, the effectiveness of impact, and the update time; where the direction of impact indicates the direction in which one running object affects another running object; the intensity of impact indicates the magnitude of the impact between running objects; the effectiveness of impact indicates whether the corresponding operational impact association information participates in operation management under the current operating condition; and the update time records the operating cycle in which the corresponding operational impact association information was last updated.

[0047] The operation management strategy generation unit first identifies the main operation impact correlation information between each operation object in the current operation cycle based on the operation impact correlation information, and determines the priority order of each operation object to participate in operation management according to the impact intensity. When multiple operation objects change their operation status at the same time, the corresponding operation object is adjusted first based on the operation impact correlation information with the greater impact intensity, and then the remaining operation objects are adjusted in turn to ensure that the operation management strategy can be consistent with the current operation conditions. The correlation priority is used to indicate the degree of priority of different operation impact correlation information when participating in the generation of operation management strategy.

[0048] Logic for calculating trade-offs in multi-association conflicts: Step 1: Extract all valid relationships within the current matrix, sort them by relationship weight from largest to smallest, and generate a scheduling priority queue; Step 2: Priority Conflict Judgment Rules: If two related relationships control the same device and the control directions are opposite, the related relationship with the larger weight value shall be taken as the main control basis; when the weight difference is less than 0.1, a compromise balance control shall be executed. Step 3: Multi-objective constraint boundary: Energy storage SOC constraint [0.2, 0.9], inverter output power not exceeding rated capacity, grid-connected power not exceeding the upper and lower limits of grid dispatch. If the boundary is exceeded, the corresponding control amount is reduced.

[0049] The correlation matrix values ​​are converted into device power command quantization rules: Input: Updated association weight matrix, current real-time operating parameters of the device; Output: Standardized digital control commands (energy storage charging and discharging power, inverter output active / reactive power, photovoltaic array start / stop regulation, grid-connected power threshold); Conversion formula: Equipment adjustment power ΔP_set = Equipment rated power × Corresponding associated weight; Example: If the photovoltaic module has a high weight of 0.8 for energy storage and the rated power of energy storage is 500kW, then the energy storage charging adjustment command ΔP_set = 500 × 0.8 = 400kW.

[0050] For example, when the ambient irradiance increases rapidly, the output capacity of photovoltaic modules changes. If the impact of photovoltaic modules on energy storage devices is stronger than the impact on grid-connected devices in the operational impact correlation information, the operation management strategy will prioritize adjusting the charging control strategy of the energy storage devices. After meeting the energy storage operation requirements, the inverter output strategy and grid-connected power control strategy will be adjusted based on the updated operational impact correlation information. When the state of charge of energy storage devices decreases, the impact of energy storage devices on grid-connected output increases in the operational impact correlation information. In this case, the operation management strategy will prioritize limiting the energy storage discharge power and simultaneously adjust the inverter output and grid-connected control strategies to ensure the continuous and stable operation of the photovoltaic power generation system.

[0051] Since the operation management strategy is always generated based on the updated operation impact association information, when the operation impact association information between operation objects changes, there is no need to rebuild the entire operation state model. Only the operation impact association information needs to be updated to achieve synchronous adjustment of the operation management strategy, thereby improving the efficiency of operation management and the real-time adaptability of the operation strategy.

[0052] The closed-loop operation management process of the management system: During the execution of the operation management strategy, the closed-loop management unit continuously obtains the operation feedback of the photovoltaic power generation system, and drives the operation impact correlation information update unit to update the operation impact correlation information in the operation status model based on the operation feedback.

[0053] Operational feedback includes one or more of the following: photovoltaic module operation feedback, inverter operation feedback, energy storage device operation feedback, grid-connected operation feedback, and environmental operation feedback. Among them, photovoltaic module operation feedback includes the actual output power, output voltage, and module temperature of the module; energy storage device operation feedback includes the state of charge, charging and discharging power, and remaining capacity; and grid-connected operation feedback includes grid-connected power, grid-connected status, and grid dispatch information.

[0054] The closed-loop operation management process of the management system: The closed-loop management unit transmits the above operation feedback to the operation impact association information update unit. The operation impact association information update unit compares the changes in operation impact association information before and after the execution of the operation management strategy and updates the operation impact association information in the operation status model.

[0055] After the update is completed, the running state model retains the running state information that has been formed in the previous running cycle, and only updates the running impact association information that has changed, so that the running state model can continuously reflect the evolution process of the running impact association information between running objects in different running cycles.

[0056] At the start of the next operating cycle, the operating management strategy generation unit directly calls the updated operating status model to generate the updated operating management strategy, so that the operating feedback formed in the previous operating cycle can continue to participate in the operating management process of the next operating cycle, realizing a closed-loop operating management mechanism of continuous updating of the operating status model, continuous optimization of operating impact correlation information, and continuous adjustment of operating management strategy.

[0057] Example 3, as Figure 6 As shown, this is the third embodiment of the present invention. Unlike the previous embodiments, this embodiment provides an artificial intelligence-based photovoltaic power generation management system to complete the management process of an operating cycle.

[0058] S1. Data Acquisition: The data acquisition unit collects real-time operational data of each operating object in the photovoltaic power generation system and sends it to the operational status modeling unit. The operational data includes photovoltaic module output voltage, output current, output power, and module temperature; inverter operating status, output power, and operating mode; energy storage equipment state of charge, charging and discharging power, and remaining capacity; grid-connected equipment operating status and grid-connected power; and parameters such as solar irradiance, ambient temperature, and wind speed collected by environmental monitoring equipment. The above operational data serves as the basic data for establishing the operational status model.

[0059] S2. Establishing an Operational Status Model: The operational status modeling unit establishes a unified operational status model based on the collected operational data. This model uniformly describes the current operational status of each operational object and stores the operational impact correlation information between them. The operational impact correlation information is not a parameter of the operational object itself, but rather information representing the influence of a change in the operational status of one operational object on the operational status of another. It includes at least the direction of influence, the intensity of influence, the effectiveness of influence, and the update time. Specifically, the direction of influence indicates the direction of transmission of the operational impact; the intensity of influence indicates the degree of operational impact; the effectiveness of influence indicates whether the corresponding operational impact participates in the current operational cycle management; and the update time records the update time of the corresponding operational impact correlation information.

[0060] The runtime sequence window is uniformly set to 15 minutes, and each sequence window is an independent runtime management cycle. The closed-loop iterative update cycle is divided into two categories: triggered update (the matrix is ​​updated immediately when the operating condition fluctuation exceeds the threshold) and periodic forced update (fixed update every 15 minutes). Runtime sequence state storage window: Saves historical runtime data for the most recent 24 hours, totaling 96 time-series cycles, as samples for incremental updates of the AI ​​model.

[0061] S3. Update Operational Impact Correlation Information: The closed-loop management unit receives operational feedback after the execution of the operational management strategy in the previous operational cycle and sends it to the operational impact correlation information update unit. The operational impact correlation information update unit compares the current operational status with the operational status of the previous operational cycle based on the operational feedback and dynamically updates the operational impact correlation information in the operational status model. When the output capacity of photovoltaic modules, the state of charge of energy storage equipment, the operating status of inverters, or the grid connection status changes, only the operational impact correlation information corresponding to the changed operational objects is updated, while the remaining operational status information remains valid. This ensures that the operational status model can continuously reflect the dynamic changes in the operational impact correlation information between operational objects as operating conditions change.

[0062] S4. Generating Operation Management Strategies: The operation management strategy generation unit reads the operation impact correlation information based on the updated operation state model, identifies the main operation impact correlation information between each operation object in the current operation cycle, determines the management priority of the operation objects according to the intensity and direction of the operation impact, and then generates corresponding operation management strategies. The operation management strategies include one or more of the following: photovoltaic module operation control strategy, energy storage charging and discharging strategy, inverter operation control strategy, grid connection control strategy, and operation and maintenance strategy. When the influence relationships between operation objects change, the operation management strategy generation unit directly adjusts the operation management strategy based on the updated operation impact correlation information without needing to rebuild the operation state model, thereby improving the efficiency of operation management response.

[0063] S5. Execute the operation management strategy and obtain operation feedback: After the operation management strategy is issued to the corresponding operating equipment, each operating object completes the corresponding control action according to the operation management strategy; the closed-loop management unit continuously collects the operation feedback after each operating object is executed. The operation feedback includes the actual power generation, energy storage charging and discharging status, inverter output status, grid connection status and environmental parameter changes, etc., and serves as the data source for updating the operation impact related information in the next operating cycle.

[0064] S6. Entering the next operating cycle: The operation impact correlation information update unit continues to update the operation impact correlation information in the operation status model based on the latest operation feedback. The updated operation status model serves as the basis for generating the operation management strategy for the next operating cycle, enabling the operation status model to be continuously saved and evolved over multiple operating cycles. The operation impact correlation information is continuously and dynamically updated according to the operating conditions, and the operation management strategy is always generated based on the latest operation status model, thus forming a photovoltaic power generation operation management mechanism driven by the operation status model, continuously updated with operation impact correlation information, and optimized in a closed loop of operation management strategy.

[0065] The foregoing has shown and described the basic principles, main features, and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited to the above embodiments. The embodiments and descriptions in the specification are merely illustrative of the principles of the invention. Various changes and modifications can be made to the invention without departing from its spirit and scope, and all such changes and modifications fall within the scope of the claimed invention.

Claims

1. A management system for photovoltaic power generation based on artificial intelligence, characterized in that, include: The data acquisition unit is used to collect operational data from photovoltaic power generation equipment, energy storage equipment, grid connection equipment, and environmental monitoring equipment in the photovoltaic power generation system. The operation status modeling unit is used to establish an operation status model of the photovoltaic power generation system based on the operation data. The operation status model is used to uniformly describe the operation status of each operation object. The operation impact correlation information update unit is used to establish operation impact correlation information between each operation object based on the operation status model, and dynamically update the operation impact correlation information based on operation feedback to reflect the changes in operation impact correlation information between each operation object. The management strategy generation unit is used to generate corresponding operation management strategies based on the operation status model and the updated operation impact correlation information. The closed-loop management unit is used to obtain the operation feedback after the operation management strategy is executed, and drive the update of the operation impact correlation information based on the operation feedback, so that the updated operation impact correlation information participates in the generation of the management strategy for the next operation cycle, thereby forming a closed-loop management of the photovoltaic power generation system driven by the operation status model.

2. The management system for photovoltaic power generation based on artificial intelligence according to claim 1, characterized in that, The operating data includes photovoltaic module output parameters, inverter operating parameters, energy storage device operating parameters, grid-connected operating parameters, and environmental monitoring parameters, wherein the environmental monitoring parameters include at least one or more of solar irradiance, ambient temperature, module temperature, and wind speed.

3. The management system for photovoltaic power generation based on artificial intelligence according to claim 1, characterized in that, The operational status model includes equipment operational status, energy transfer status, and operational sequence status, which are used to describe the current operational status of each operational object and its corresponding operational attributes.

4. The management system for photovoltaic power generation based on artificial intelligence according to claim 3, characterized in that, The operational impact-related information includes at least one or more of the following: power transfer relationships, energy exchange relationships, control response relationships, and scheduling dependencies.

5. A management system for photovoltaic power generation based on artificial intelligence according to claim 4, characterized in that, The operation impact association information update unit dynamically adjusts the operation impact association information based on operation feedback. The dynamic adjustment includes at least one or more of the following: association weight update, association direction update, association validity update, and association priority update.

6. The management system for photovoltaic power generation based on artificial intelligence according to claim 1, characterized in that, The management strategy generation unit generates at least one of the following based on the operating status model and operating impact correlation information: photovoltaic array operation mode, energy storage charging and discharging strategy, inverter control strategy, grid connection control strategy, and operation and maintenance strategy.

7. A management system for photovoltaic power generation based on artificial intelligence according to claim 1, characterized in that, The closed-loop management unit continuously acquires operational feedback and forms a feedback data stream.

8. A management system for photovoltaic power generation based on artificial intelligence according to claim 7, characterized in that, The operational feedback includes one or more of the following: power generation change information, energy storage status change information, grid connection status change information, equipment response information, and environmental change information.

9. A management system for photovoltaic power generation based on artificial intelligence according to claim 1, characterized in that, After the operation impact association information is updated, it re-participates in the operation management strategy generation to ensure that the newly generated operation management strategy is consistent with the current operation impact association information.

10. A management system for photovoltaic power generation based on artificial intelligence according to claim 1, characterized in that, The operational impact information is organized using an association matrix data structure.

Images