Photovoltaic power station control optimization method and system

Abstract

The invention provides a photovoltaic power station control optimization method and system, and belongs to the field of photovoltaic power stations. The method comprises the following steps: collecting multi-source environment data, and extracting space-time correlation characteristics; generating a shadow thermodynamic diagram gray matrix by using a space-time GAN model; calculating control parameters based on the shadow thermodynamic diagram gray scale matrix, and selecting a control strategy by using a bimodal mapping function to control the photovoltaic power station; converting the control parameter into an impedance coefficient, and embedding the impedance coefficient into a grid-connected inverter control value function so as to suppress power grid harmonic oscillation of the photovoltaic power station; and calculating an efficiency index to optimize the bimodal mapping function, and calibrating parameters of the space-time GAN model to optimize control of the photovoltaic power station. According to the method, the change trend of the shadow coverage rate can be pre-judged, control lag is prevented, and different control strategies are selected for the photovoltaic power station; the control response speed is improved, the power grid harmonic oscillation of the photovoltaic power station is suppressed, and the power mutation is suppressed.

Description

Technical Field

[0001] This invention relates to the field of photovoltaic power plant technology, and more specifically to a photovoltaic power plant control optimization method and system. Background Technology

[0002] A photovoltaic (PV) power station refers to a power generation system that utilizes solar energy and employs electronic components such as crystalline silicon panels and grid-connected inverters. With increasing energy demand and growing environmental concerns, PV power generation, as an important component of clean energy, is being used more and more widely. However, due to the wide distribution, numerous connection points, and complex operating environments of PV power stations, their control and optimization methods face many challenges.

[0003] In existing technologies, photovoltaic power plant control technology mainly relies on traditional PID control systems and data-driven optimization control. However, these control technologies still have the following drawbacks: traditional PID control systems rely on fixed parameters, have insufficient dynamic response capabilities, and cannot predict environmental change trends, resulting in control lag; existing harmonic monitoring and control strategies are separated, lacking feedback adjustment mechanisms, resulting in governance lagging behind disturbances; moreover, current prediction models have large errors in predicting weather fluctuations, resulting in photovoltaic power plant operation and maintenance strategies lacking predictability and economy. Summary of the Invention

[0004] The purpose of this invention is to provide a control optimization method and system for photovoltaic power plants, which solves the problems in the existing photovoltaic power plant control technology that cannot predict environmental change trends, resulting in control lag, separation of harmonic monitoring and control strategies, lack of feedback adjustment mechanisms, and lack of predictability and economy in photovoltaic power plant operation and maintenance strategies.

[0005] To achieve the above objectives, this invention provides a control optimization method for photovoltaic power plants. The method includes: collecting multi-source environmental data based on sensors deployed at the photovoltaic power plant and extracting spatiotemporal correlation features; generating a shadow heatmap grayscale matrix using a spatiotemporal GAN ​​model based on the spatiotemporal correlation features; calculating control parameters based on the shadow heatmap grayscale matrix and selecting a control strategy using a bimodal mapping function to control the photovoltaic power plant; converting the control parameters into impedance coefficients and embedding the impedance coefficients into the grid-connected inverter control value function to suppress grid harmonic oscillations of the photovoltaic power plant; calculating an efficiency index based on the harmonic distortion rate of the photovoltaic power plant and the impedance coefficients; optimizing the bimodal mapping function using the efficiency index; and calibrating the parameters of the spatiotemporal GAN ​​model using the optimized bimodal mapping function to optimize the control of the photovoltaic power plant.

[0006] Optionally, the step of generating a grayscale matrix of a shadow heatmap using a spatiotemporal GAN ​​model includes: using the generator of the spatiotemporal GAN ​​model to extract and fuse shallow and deep features of spatiotemporal correlation features, and outputting an initial grayscale matrix of the heatmap; using the discriminator of the spatiotemporal GAN ​​model to compare the feature similarity between the initial grayscale matrix of the heatmap and the shadow distribution of the photovoltaic power station; the generator updates the weights according to the feature similarity feedback from the discriminator, so that the discriminator cannot distinguish the features of the initial grayscale matrix of the heatmap and the shadow distribution of the photovoltaic power station, and outputs the grayscale matrix of the shadow heatmap.

[0007] Optionally, calculating the control parameters based on the grayscale matrix of the shadow heatmap includes: calculating the grayscale mean value based on the grayscale matrix of the shadow heatmap. Gray-scale difference value The control parameters are obtained, wherein each element in the grayscale matrix of the shadow heatmap is... Indicates the first OK The predicted shadow coverage intensity of the column component, the mean grayscale value The grayscale difference value represents the average shadow coverage intensity in the grayscale matrix of the shadow heatmap. The spatial migration rate of shadow coverage per unit time.

[0008] Optionally, the step of using a dual-mode mapping function to select a control strategy includes: if the volatility of the grayscale mean does not exceed a volatility threshold, then a steady-state mapping mode is selected to control the photovoltaic power station; if the grayscale difference value is greater than a difference value threshold, then a transient mapping mode is selected to control the photovoltaic power station.

[0009] Optionally, the dual-mode mapping function includes a steady-state mapping function and a transient mapping function, and the step of converting the control parameter into an impedance coefficient includes: if the steady-state mapping mode is selected, then the steady-state mapping function is used to convert the gray-scale mean value into an impedance coefficient; if the transient mapping mode is selected, then the transient mapping function is used to convert the gray-scale difference value into an impedance coefficient.

[0010] Optionally, the step of using the grid-connected inverter control value function to suppress grid harmonic oscillations of the photovoltaic power station includes: determining the rate of change of shadow coverage intensity based on the gray-scale matrix of the shadow heat map; maintaining the steady-state mapping mode when the rate of change of shadow coverage intensity does not exceed a preset threshold; and activating the transient mapping mode when the rate of change of shadow coverage intensity exceeds the preset threshold, thereby increasing the impedance coefficient and enhancing the weight of the disturbance suppression term in the grid-connected inverter control value function to suppress grid harmonic oscillations of the photovoltaic power station.

[0011] Optionally, the calculation of efficiency indicators based on the harmonic distortion rate of the photovoltaic power station and the impedance coefficient includes: collecting voltage signals based on monitoring equipment deployed at the photovoltaic power station, and converting the voltage signals into frequency domain signals through fast Fourier transform; extracting harmonic parameters based on the frequency domain signals, and calculating the harmonic distortion rate of the photovoltaic power station; when the negative sequence voltage characteristic value of the power grid in the gray-scale matrix of the shadow heat map exceeds the limit, the efficiency indicators are calculated based on the harmonic distortion rate and the impedance coefficient.

[0012] Optionally, optimizing the bimodal mapping function using the efficiency index includes: if the efficiency index does not exceed a benchmark value within a continuous preset period, optimizing the steady-state mapping function to enhance the weak shadow response, and optimizing the transient mapping function to suppress overcompensation.

[0013] Optionally, calibrating the parameters of the spatiotemporal GAN ​​model using the optimized bimodal mapping function includes: generating a corrected impedance coefficient using the optimized bimodal mapping function; and generating an auxiliary loss term for the discriminator of the spatiotemporal GAN ​​model based on the corrected impedance coefficient, so as to force the generator of the spatiotemporal GAN ​​model to optimize the parameters of the spatiotemporal GAN ​​model.

[0014] On the other hand, the present invention provides a photovoltaic power station control optimization system, the control optimization system including a control module, the control module including a memory, a processor and a computer program stored in the memory and executable on the processor, the processor executing the computer program to implement the control optimization method described above.

[0015] Through the above technical solutions, this invention generates a component-level shadow heatmap for a future period using a spatiotemporal GAN ​​model, predicts the trend of shadow coverage changes, and prevents control lag. By using the grayscale matrix of the shadow heatmap and the dual-modal mapping function, the predicted results directly drive the control parameters, allowing for the selection of different control strategies for photovoltaic power plants, improving the control response speed, and enabling adaptation to second-level shadow migration. By adding a shadow disturbance penalty term to the control value function of the grid-connected inverter, the grid harmonic oscillation of the photovoltaic power plant can be suppressed, and power surges can be inhibited.

[0016] Other features and advantages of the embodiments of the present invention will be described in detail in the following detailed description section. Attached Figure Description

[0017] The accompanying drawings are provided to further illustrate embodiments of the present invention and form part of the specification. They are used together with the following detailed description to explain the embodiments of the present invention, but do not constitute a limitation thereof. In the drawings: Figure 1This is a flowchart illustrating a photovoltaic power plant control optimization method according to the present invention. Figure 1 ; Figure 2 This is a schematic diagram of the process for generating the grayscale matrix of the shadow heatmap in this invention; Figure 3 This is a flowchart illustrating the calculation of efficiency indicators in this invention; Figure 4 This is a flowchart illustrating a photovoltaic power plant control optimization method according to the present invention. Figure 2 . Detailed Implementation

[0018] The specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings. It should be understood that the specific embodiments described herein are for illustration and explanation only and are not intended to limit the scope of the present invention.

[0019] It should be noted that the acquisition, transmission, storage, use, and processing of data in the technical solution of this application all comply with the relevant provisions of national laws and regulations. In the embodiments of this application, certain existing industry solutions such as software, components, and models may be mentioned. These should be considered exemplary, intended only to illustrate the feasibility of implementing the technical solution of this application, and do not imply that the applicant has already used or necessarily used such solutions.

[0020] First embodiment: Please refer to Figure 1 and Figure 4 This invention provides a control optimization method for a photovoltaic power station, which may include: Step S110: Based on the sensors deployed at the photovoltaic power station, collect multi-source environmental data and extract spatiotemporal correlation features.

[0021] In this embodiment of the invention, the sensor may include a meteorological sensor, a component-level current sensor, and a grid-connected point harmonic analyzer, etc. The multi-source environmental data may include cloud motion vectors (e.g., cloud orientation angle or velocity) and irradiance time series data collected by the meteorological sensor, the component-level current sensor captures current gradient characteristics in real time and marks power change timestamps, and the grid-connected point harmonic analyzer collects negative sequence voltage component characteristics.

[0022] In a preferred embodiment of the present invention, cloud orientation angle features, current gradient features and negative sequence voltage component features can be fused into spatiotemporal correlation features. Cloud orientation angle features can be used to predict cloud trajectory modeling, current gradient features can be used as a power mutation verification benchmark, and negative sequence voltage component features can be used to perform layer benchmark value calibration.

[0023] Step S120: Based on the spatiotemporal correlation features, use the spatiotemporal GAN ​​model to generate a grayscale matrix of the shadow heatmap.

[0024] Please refer to Figure 2 In a preferred embodiment of the present invention, generating a grayscale matrix of a shadow heatmap using a spatiotemporal GAN ​​model may include steps S1201-S1203: Step S1201: Using the generator of the spatiotemporal GAN ​​model, extract and fuse the shallow and deep features of the spatiotemporal correlation features, and output the initial heatmap grayscale matrix.

[0025] The spatiotemporal GAN ​​model consists of a generator and a discriminator. The shallow features extracted by the generator can refer to the spatial texture (e.g., cloud edges) initially extracted by the convolutional layer, while the deep features can be cloud displacement and cross-regional cloud migration patterns at times T1→T2→T3. The initial heatmap grayscale matrix can be output by fusing the shallow and deep features.

[0026] Step S1202: Use the discriminator of the spatiotemporal GAN ​​model to compare the feature similarity between the gray-scale matrix of the initial heat map and the shadow distribution of the photovoltaic power station.

[0027] The discriminator of the spatiotemporal GAN ​​model can include a temporal consistency discriminator, a spatial continuity discriminator, and a spectral authenticity discriminator. The temporal consistency discriminator can ensure that the generated results of the spatiotemporal GAN ​​model conform to the historical cloud movement trend (e.g., migration speed stability) by using the shadow change difference at time T1 to T2. The spatial continuity discriminator can verify the spatial occlusion continuity (e.g., the naturalness of shadow edge transition) by using the shadow change difference at time T2 to T3. The spectral authenticity discriminator can constrain spectral fidelity (e.g., the matching of gray intensity with the actual cloud thickness) by comparing the result of the shadow heatmap transformed by the spectral response function with the actual multispectral image.

[0028] Step S1203: The generator updates the weights based on the feature similarity feedback from the discriminator, so that the discriminator cannot distinguish the features of the initial heat map grayscale matrix and the shadow distribution of the photovoltaic power station, and outputs the shadow heat map grayscale matrix.

[0029] Each matrix element in the grayscale matrix of the shadow heatmap Indicates the first OK The predicted shadow coverage intensity of the column component has a value range of [0, 1]. When the predicted shadow coverage intensity is 0, it means there is no shadow coverage, i.e., a fully illuminated area. When the predicted shadow coverage intensity is 1, it means complete occlusion, such as a cloud-covered area.

[0030] This invention uses a spatiotemporal GAN ​​model to generate a component-level shadow heatmap for a future period of time (e.g., the next 5 minutes), predicts the trend of shadow coverage changes, and prevents control lag.

[0031] Step S130: Based on the grayscale matrix of the shadow heat map, calculate the control parameters and use the dual-mode mapping function to select the control strategy to control the photovoltaic power station.

[0032] In a preferred embodiment of the present invention, calculating control parameters based on the grayscale matrix of the shadow heatmap may include: calculating the mean grayscale value based on the grayscale matrix of the shadow heatmap. Gray-scale difference value The control parameters are obtained, including the mean gray level. The grayscale difference value represents the average shadow coverage intensity in the grayscale matrix of the shadow heatmap. The spatial migration rate of shadow coverage per unit time.

[0033] In a preferred embodiment of the present invention, the grayscale mean can be represented by the following formula. :

[0034] in, Indicates the first Line number The grayscale values ​​of the pixels in the column matrix. Indicates the number of rows in the matrix. Indicates the number of columns in the matrix.

[0035] In a preferred embodiment of the present invention, the grayscale difference value can be expressed by the following formula. :

[0036] in, Indicates the first Line number Horizontal difference of column matrix pixels Indicates the first Line number Vertical difference of column matrix pixels. This indicates the time interval between two adjacent frames.

[0037] In a preferred embodiment of the present invention, the control strategy selected using the dual-modal mapping function may include: if the volatility of the grayscale mean does not exceed a volatility threshold (e.g., 12%), then a steady-state mapping mode is selected to control the photovoltaic power station. The steady-state mapping mode can maximize the energy capture efficiency of the photovoltaic array and maintain smooth grid-connected power output under stable light and temperature conditions; if the grayscale difference value is greater than a difference threshold (e.g., 18 grayscale / second), then a transient mapping mode is selected to control the photovoltaic power station. The transient mapping mode can suppress power oscillations and maintain grid-connected stability during grid faults, load changes, or rapid cloud migration (e.g., shadow change rate > 20 grayscale / second).

[0038] Step S140: Convert the control parameters into impedance coefficients and embed the impedance coefficients into the grid-connected inverter control value function to suppress grid harmonic oscillations of the photovoltaic power station using the grid-connected inverter control value function.

[0039] In a preferred embodiment of the present invention, the dual-mode mapping function includes a steady-state mapping function and a transient mapping function. Converting control parameters into impedance coefficients may include: if the steady-state mapping mode is selected, the steady-state mapping function is used to convert the gray-scale mean into impedance coefficients for proportional-integral control. In this case, control accuracy is prioritized, and the impedance coefficient reflects the system's suppression strength against shadow disturbances. A high impedance coefficient indicates a stronger disturbance suppression capability.

[0040] The impedance coefficient calculated using the steady-state mapping function can be expressed by the following formula. :

[0041] Furthermore, if the transient mapping mode is selected, the transient mapping function is used to convert the grayscale difference value into an impedance coefficient for pure proportional control, in which case the control speed takes priority.

[0042] The impedance coefficient calculated using the transient mapping function can be expressed by the following formula. :

[0043] in, This is the attenuation coefficient.

[0044] In a preferred embodiment of the present invention, suppressing grid harmonic oscillations of a photovoltaic power station using a grid-connected inverter control value function may include: determining the rate of change of shadow coverage intensity based on the gray-scale matrix of the shadow heat map; maintaining a steady-state mapping mode when the rate of change of shadow coverage intensity does not exceed a preset threshold (e.g., 12%); activating a transient mapping mode when the rate of change of shadow coverage intensity exceeds the preset threshold, indicating that the shadow coverage rate is accelerating, increasing the impedance coefficient, enhancing the weight of the disturbance suppression term in the grid-connected inverter control value function, automatically pre-activating the low-voltage ride-through mode, injecting negative sequence current, and switching to pure proportional mode control to suppress grid harmonic oscillations of the photovoltaic power station.

[0045] The control value function of a grid-connected inverter can be expressed by the following formula. :

[0046] in, This represents the reference value for active power. Indicates the actual active power. This represents the reference value for reactive power. Indicates actual reactive power. This represents the reactive power weighting coefficient. This represents the perturbation suppression weighting coefficient. Indicates the impedance coefficient. This represents the disturbance suppression term.

[0047] Step S150: Based on the harmonic distortion rate and impedance coefficient of the photovoltaic power station, calculate the efficiency index to optimize the dual-mode mapping function through the efficiency index, and use the optimized dual-mode mapping function to calibrate the parameters of the spatiotemporal GAN ​​model to optimize the control of the photovoltaic power station.

[0048] Please refer to Figure 3 In a preferred embodiment of the present invention, calculating the efficiency index based on the harmonic distortion rate and impedance coefficient of the photovoltaic power station may include steps S1501-S1503: Step S1501: The monitoring equipment deployed based on the photovoltaic power station collects the voltage signal and converts the voltage signal into a frequency domain signal through fast Fourier transform.

[0049] Step S1502: Based on the frequency domain signal, extract the harmonic parameters and calculate the harmonic distortion rate of the photovoltaic power station.

[0050] In a preferred embodiment of the present invention, after the acquired voltage signal is converted into a frequency domain signal by fast Fourier transform, the fundamental frequency (50Hz) and various harmonic components (e.g., 3rd, 5th, 7th, etc.) can be decomposed, and the effective value of the fundamental frequency can be calculated. With the effective values ​​of each harmonic (For example, the 3rd (150Hz) and 5th (250Hz) are used as harmonic parameters to calculate the harmonic distortion rate of photovoltaic power plants. The higher the harmonic distortion rate, the more severe the waveform distortion, and the lower the harmonic distortion rate, the better the power quality.

[0051] The harmonic distortion rate of a photovoltaic power station can be expressed by the following formula. :

[0052] in, Indicates the effective value of the fundamental voltage. Indicates the first Effective value of subharmonic voltage.

[0053] Step S1503: When the negative sequence voltage characteristic value of the power grid in the gray-scale matrix of the shaded heat map exceeds the limit, the efficiency index is calculated based on the harmonic distortion rate and impedance coefficient.

[0054] In a preferred embodiment of the present invention, the efficiency index can be represented by the following formula. :

[0055] in, Indicates the impedance coefficient. Indicates harmonic distortion rate. A higher value indicates better control efficiency, meaning the system achieves disturbance suppression at a lower harmonic cost.

[0056] In a preferred embodiment of the present invention, optimizing the bimodal mapping function through an efficiency index may include: if the efficiency index does not exceed a benchmark value (e.g., 0.8) within a continuous preset period, then optimizing the steady-state mapping function to enhance the weak shadow response, and optimizing the transient mapping function to suppress overcompensation.

[0057] In a preferred embodiment of the present invention, the steady-state mapping function is used at low gray levels (e.g., The response is smooth and the suppression under weak shadows is insufficient. Therefore, the steady-state mapping function is optimized to linearly enhance the sensitivity of the impedance coefficient in the low grayscale region, so as to improve the power maintenance capability under weak shadows.

[0058] The impedance coefficient calculated using the optimized steady-state mapping function can be expressed by the following formula. :

[0059] In a preferred embodiment of the present invention, the transient mapping function is applied in high-speed cloud layers (e.g., When the impedance coefficient increases dramatically during high-speed disturbances, it leads to overcompensation. Therefore, the transient mapping function is optimized to suppress the impedance coefficient growth rate under high-speed disturbances and avoid overcharging or over-discharging of energy storage devices.

[0060] The impedance coefficient calculated using the optimized transient mapping function can be expressed by the following formula. :

[0061] In a preferred embodiment of the present invention, calibrating the parameters of the spatiotemporal GAN ​​model using an optimized bimodal mapping function may include: generating a corrected impedance coefficient using the optimized bimodal mapping function; generating an auxiliary loss term for the discriminator of the spatiotemporal GAN ​​model based on the corrected impedance coefficient, so as to force the generator of the spatiotemporal GAN ​​model to optimize the parameters of the spatiotemporal GAN ​​model; and injecting the corrected impedance coefficient as prior knowledge into the spatiotemporal GAN ​​model to improve the prediction accuracy of the impedance coefficient.

[0062] In a preferred embodiment of the present invention, generating an auxiliary loss term for the discriminator of the spatiotemporal GAN ​​model to force the generator of the spatiotemporal GAN ​​model to optimize the parameters of the spatiotemporal GAN ​​model may include steps S2501-S2503: Step S2501: Add shadow features ( The generator takes the spatiotemporal GAN ​​model as input and outputs the predicted impedance coefficients. The true impedance coefficient or Input the discriminator of the spatiotemporal GAN ​​model to determine the authenticity of the data.

[0063] Step S2502: Based on the predicted impedance coefficient, the true impedance coefficient, and the adversarial loss, design an auxiliary loss function to generate the auxiliary loss term of the discriminator of the spatiotemporal GAN ​​model.

[0064] The auxiliary loss term can be expressed by the following formula. :

[0065] in, This represents an adversarial loss mechanism, ensuring that the generated data distribution approximates the true distribution. This represents the weighting coefficient (usually between 0.5 and 1.0), balancing adversarial losses and physical consistency. The spatiotemporal GAN ​​model generator is forced to learn the corrected impedance coefficient mapping relationship.

[0066] Step S2503: Optimization objective for training the spatiotemporal GAN ​​model generator The optimization objective of the discriminator in the training spatiotemporal GAN ​​model .

[0067] Step S2504: Input real-time shadow features, the generator outputs the predicted impedance coefficient, the control system uses the predicted impedance coefficient to calculate the efficiency index, if the efficiency index continues to be lower than the limit again, a new round of bimodal mapping function optimization is triggered, and the spatiotemporal GAN ​​model training dataset is updated.

[0068] Accordingly, this invention provides a photovoltaic power plant control optimization method, which includes: collecting multi-source environmental data based on sensors deployed at the photovoltaic power plant and extracting spatiotemporal correlation features; generating a shadow heatmap grayscale matrix based on the spatiotemporal correlation features using a spatiotemporal GAN ​​model; calculating control parameters based on the shadow heatmap grayscale matrix and selecting a control strategy using a dual-mode mapping function to control the photovoltaic power plant; converting the control parameters into impedance coefficients and embedding the impedance coefficients into the grid-connected inverter control value function to suppress grid harmonic oscillations of the photovoltaic power plant; calculating an efficiency index based on the harmonic distortion rate and impedance coefficients of the photovoltaic power plant to optimize the dual-mode mapping function through the efficiency index, and calibrating the parameters of the spatiotemporal GAN ​​model using the optimized dual-mode mapping function to optimize the control of the photovoltaic power plant. This invention generates a component-level shadow heatmap for a future period using a spatiotemporal GAN ​​model, predicting the trend of shadow coverage changes and preventing control lag. By using the grayscale matrix of the shadow heatmap and a dual-modal mapping function, the predicted results directly drive control parameters, allowing for the selection of different control strategies for photovoltaic power plants, improving control response speed, and adapting to second-level shadow migration. By adding a shadow disturbance penalty term to the control value function of the grid-connected inverter, grid harmonic oscillations of photovoltaic power plants can be suppressed, and power surges can be inhibited.

[0069] This invention also provides a photovoltaic power station control optimization system. The control optimization system includes a control module, which includes a memory, a processor, and a computer program stored in the memory and executable on the processor. The processor executes the computer program to implement the control optimization method described above.

[0070] Those skilled in the art will understand that embodiments of this application can be provided as methods, systems, or computer program products. Therefore, this application can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, this application can take the form of a computer program product embodied on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0071] This application is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this application. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart... Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0072] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0073] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0074] In a typical configuration, a computing device includes one or more processors (CPU), input / output interfaces, network interfaces, and memory.

[0075] Memory may include non-persistent memory in computer-readable media, such as random access memory (RAM) and / or non-volatile memory, like read-only memory (ROM) or flash RAM. Memory is an example of computer-readable media.

[0076] Computer-readable media includes both permanent and non-permanent, removable and non-removable media that can store information using any method or technology. Information can be computer-readable instructions, data structures, modules of programs, or other data. Examples of computer storage media include, but are not limited to, phase-change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technologies, CD-ROM, digital versatile optical disc (DVD) or other optical storage, magnetic tape, magnetic disk storage or other magnetic storage devices, or any other non-transferable medium that can be used to store information accessible by a computing device. As defined herein, computer-readable media does not include transient computer-readable media, such as modulated data signals and carrier waves.

[0077] It should also be noted that the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article, or apparatus. Unless otherwise specified, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes that element.

[0078] The above are merely embodiments of this application and are not intended to limit the scope of this application. Various modifications and variations can be made to this application by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the scope of the claims of this application.

Claims

1. A method for optimizing the control of a photovoltaic power plant, characterized in that, The control optimization method includes: Sensors deployed at photovoltaic power plants are used to collect multi-source environmental data and extract spatiotemporal correlation features. Based on the aforementioned spatiotemporal correlation features, a grayscale matrix of shadow heatmap is generated using a spatiotemporal GAN ​​model. Based on the grayscale matrix of the shadow heat map, control parameters are calculated, and a control strategy is selected using a bimodal mapping function to control the photovoltaic power station. The control parameters are converted into impedance coefficients, and the impedance coefficients are embedded into the grid-connected inverter control value function to suppress grid harmonic oscillations of the photovoltaic power station using the grid-connected inverter control value function. Based on the harmonic distortion rate and impedance coefficient of the photovoltaic power station, an efficiency index is calculated to optimize the bimodal mapping function. The optimized bimodal mapping function is then used to calibrate the parameters of the spatiotemporal GAN ​​model to optimize the control of the photovoltaic power station.

2. The control optimization method according to claim 1, characterized in that, The process of generating a grayscale matrix for a shadow heatmap using a spatiotemporal GAN ​​model includes: Using the generator of the spatiotemporal GAN ​​model, shallow and deep features of spatiotemporal correlation features are extracted and fused to output the initial heatmap grayscale matrix. The discriminator of the spatiotemporal GAN ​​model is used to compare the feature similarity between the gray-scale matrix of the initial heat map and the shadow distribution of the photovoltaic power station. The generator updates the weights based on the feature similarity feedback from the discriminator, so that the discriminator cannot distinguish the features of the initial heat map grayscale matrix from the shadow distribution of the photovoltaic power station, and outputs the shadow heat map grayscale matrix.

3. The control optimization method according to claim 1, characterized in that, The calculation of control parameters based on the grayscale matrix of the shadow heatmap includes: Based on the grayscale matrix of the shadow heatmap, the grayscale mean is calculated respectively. Gray-scale difference value To obtain the control parameters, Each element in the grayscale matrix of the shadow heatmap Indicates the first OK The predicted shadow coverage intensity of the column component, the mean grayscale value The grayscale difference value represents the average shadow coverage intensity in the grayscale matrix of the shadow heatmap. The spatial migration rate of shadow coverage per unit time.

4. The control optimization method according to claim 1, characterized in that, The selection of control strategy using a bimodal mapping function includes: If the volatility of the grayscale mean does not exceed the volatility threshold, then the steady-state mapping mode is selected to control the photovoltaic power station. If the grayscale difference value is greater than the difference value threshold, then the transient mapping mode is selected to control the photovoltaic power station.

5. The control optimization method according to claim 4, characterized in that, The dual-mode mapping function includes a steady-state mapping function and a transient mapping function. Converting the control parameters into impedance coefficients includes: If the steady-state mapping mode is selected, the steady-state mapping function is used to convert the grayscale mean into the impedance coefficient. If the transient mapping mode is selected, the transient mapping function is used to convert the grayscale difference value into an impedance coefficient.

6. The control optimization method according to claim 1, characterized in that, The method of using the grid-connected inverter to control the value function and suppress grid harmonic oscillations of the photovoltaic power station includes: Based on the grayscale matrix of the shadow heatmap, determine the rate of change of shadow coverage intensity; When the rate of change of shadow coverage intensity does not exceed a preset threshold, the steady-state mapping mode is maintained; When the rate of change of the shadow coverage intensity exceeds a preset threshold, the transient mapping mode is activated to increase the impedance coefficient and enhance the weight of the disturbance suppression term in the control value function of the grid-connected inverter, so as to suppress the grid harmonic oscillation of the photovoltaic power station.

7. The control optimization method according to claim 1, characterized in that, The efficiency index is calculated based on the harmonic distortion rate and impedance coefficient of the photovoltaic power station, including: The monitoring equipment deployed in the photovoltaic power station collects voltage signals and converts the voltage signals into frequency domain signals through fast Fourier transform; Based on the frequency domain signal, harmonic parameters are extracted, and the harmonic distortion rate of the photovoltaic power station is calculated. When the negative sequence voltage characteristic value of the power grid in the gray-scale matrix of the shadow heatmap exceeds the limit, the efficiency index is calculated based on the harmonic distortion rate and the impedance coefficient.

8. The control optimization method according to claim 5, characterized in that, The optimization of the bimodal mapping function using the efficiency index includes: If the efficiency index does not exceed the benchmark value within a continuous preset period, the steady-state mapping function is optimized to enhance the weak shadow response, and the transient mapping function is optimized to suppress overcompensation.

9. The control optimization method according to claim 1, characterized in that, The calibration of the spatiotemporal GAN ​​model parameters using an optimized bimodal mapping function includes: The corrected impedance coefficient is generated using an optimized dual-mode mapping function; Based on the corrected impedance coefficient, an auxiliary loss term is generated for the discriminator of the spatiotemporal GAN ​​model, thereby forcing the generator of the spatiotemporal GAN ​​model to optimize the parameters of the spatiotemporal GAN ​​model.

10. A photovoltaic power plant control optimization system, characterized in that, The control optimization system includes a control module, which includes a memory, a processor, and a computer program stored in the memory and executable on the processor. The processor executes the computer program to implement the control optimization method according to any one of claims 1-9.

Images