Flexible adjustable control method and device for distributed photovoltaic system

Abstract

The invention provides a flexible adjustable control method and device for a distributed photovoltaic system, and relates to the technical field of power network intelligent scheduling, and the method comprises the steps: obtaining multi-mode sensing time sequence data of a deployed photovoltaic array in a historical preset time period and weather forecast data corresponding to a future preset time period through an edge node; predicting the predicted photovoltaic generating capacity and predicted power supply load demand of the corresponding photovoltaic array based on the time sequence prediction module; predicting an initial scheduling parameter of the corresponding photovoltaic array based on the scheduling decision model; sending the predicted photovoltaic power generation amount, the predicted power supply load demand and the initial scheduling parameter to a core node, so that the core node updates a target scheduling parameter for each edge node; and receiving the corresponding target scheduling parameter from the core node, and regulating and controlling the corresponding photovoltaic array according to the target scheduling parameter. Therefore, the global optimization dynamic cooperative control method is implemented in the large-scale distributed photovoltaic system, and the reliability of global flexible scheduling is improved.

Description

Technical Field

[0001] The present application relates to the technical field of intelligent dispatching of power networks, and in particular to a flexible and adjustable control method and device for a distributed photovoltaic system. Background Art

[0002] As the global demand for clean energy grows, distributed photovoltaic systems have been widely used as an important form of renewable energy. Through distributed photovoltaic systems, solar energy resources can be used to directly generate electricity at or near the user and provide electricity to the grid. However, with the large-scale deployment of distributed photovoltaic systems, some technical challenges have gradually emerged.

[0003] At present, most distributed photovoltaic systems adopt a control strategy based on fixed parameters. Although a variety of typical working conditions are considered in the early design, in actual applications, due to the volatility of solar energy resources and the uncertainty of user load demand, the output of the photovoltaic system often needs to be dynamically adjusted. However, the existing fixed parameter control strategy cannot effectively cope with these changes, resulting in frequent mismatches between the output of the photovoltaic system and the grid demand. This control rigidity not only reduces the power generation efficiency of the photovoltaic system, but also increases the difficulty of grid dispatching.

[0004] In response to the above problems, the industry has not yet proposed a better technical solution. Summary of the invention

[0005] The present application provides a flexible and adjustable control method, device, storage medium, computer program product and electronic device for a distributed photovoltaic system, which is used to at least solve the problem of mismatch between the output of the photovoltaic system and the demand of the power grid caused by the rigidity of the current photovoltaic system control.

[0006] In a first aspect, an embodiment of the present application provides a flexible and adjustable control method for a distributed photovoltaic system, wherein the distributed photovoltaic system includes multiple photovoltaic arrays, and corresponding sensor networks and edge nodes are deployed for each photovoltaic array, respectively. The method includes: each edge node is used to perform the following operations: obtaining multimodal sensing time series data of the deployed photovoltaic array in a historical preset time period and weather forecast data corresponding to a future preset time period; the multimodal sensing time series data includes multiple adjacent historical time steps and corresponding multimodal sensing data, and the parameter types of the multimodal sensing data include: solar radiation intensity, ambient temperature, power generation, grid operation parameters and power supply load demand; inputting the multimodal sensing time series data and the weather forecast data into a time series prediction model to predict the corresponding photovoltaic array in the future. The estimated photovoltaic power generation and the estimated power supply load demand in a preset time period; the time series prediction model adopts an LSTM (Long Short-Term Memory) model; the corresponding estimated photovoltaic power generation and the estimated power supply load demand are processed based on the scheduling decision model to predict the initial scheduling parameters of the corresponding photovoltaic array; the parameter type of the scheduling parameter includes the target output power and the grid-connected voltage adjustment value; the estimated photovoltaic power generation, the estimated power supply load demand and the initial scheduling parameters are sent to the core node, so that the core node performs a global analysis on the data sent by each edge node to generate a target scheduling parameter for each edge node; the corresponding target scheduling parameter is received from the core node, and the corresponding photovoltaic array is adjusted in the future preset time period according to the target scheduling parameter.

[0007] In a second aspect, an embodiment of the present application provides a flexible and adjustable control device for a distributed photovoltaic system, wherein the distributed photovoltaic system includes multiple photovoltaic arrays, and corresponding sensor networks and edge nodes are deployed for each photovoltaic array, respectively. The device includes an acquisition unit, a supply and demand prediction unit, an initial scheduling prediction unit, a data upload unit, and a target scheduling control unit arranged in the edge node; the acquisition unit is used to acquire multimodal sensing time series data of the deployed photovoltaic array in a historical preset time period and weather forecast data corresponding to a future preset time period; the multimodal sensing time series data includes multiple adjacent historical time steps and corresponding multimodal sensing data, and the parameter types of the multimodal sensing data include: solar radiation intensity, ambient temperature, power generation, power grid operation parameters, and power supply load demand; the supply and demand prediction unit is used to input the multimodal sensing time series data and the weather forecast data into a time series data frame; A time series prediction model is used to predict the expected photovoltaic power generation and the expected power supply load demand of the corresponding photovoltaic array in the future preset time period; the time series prediction model adopts an LSTM model; an initial scheduling prediction unit is used to process the corresponding expected photovoltaic power generation and the expected power supply load demand based on the scheduling decision model to predict the initial scheduling parameters of the corresponding photovoltaic array; the parameter types of the scheduling parameters include target output power and grid-connected voltage adjustment value; a data uploading unit is used to send the expected photovoltaic power generation, the expected power supply load demand and the initial scheduling parameters to the core node, so that the core node performs a global analysis on the data sent by each edge node to generate a target scheduling parameter for each edge node; a target scheduling control unit is used to receive the corresponding target scheduling parameter from the core node, and regulate the corresponding photovoltaic array in the future preset time period according to the target scheduling parameter.

[0008] According to a third aspect, an electronic device is provided, comprising: at least one processor, and a memory communicatively connected to the at least one processor, wherein the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor so that the at least one processor can perform the steps of the flexible and adjustable control method of a distributed photovoltaic system of any embodiment of the present application.

[0009] In a fourth aspect, an embodiment of the present application provides a storage medium on which a computer program is stored, characterized in that when the program is executed by a processor, the steps of the flexible and adjustable control method of a distributed photovoltaic system of any embodiment of the present application are implemented.

[0010] In a fifth aspect, an embodiment of the present application provides a computer program product, including a computer program / instruction, which, when executed by a processor, implements the steps of the flexible and adjustable control method of a distributed photovoltaic system of any embodiment of the present application.

[0011] The flexible and adjustable control method of a distributed photovoltaic system provided by the present application can produce at least the following technical effects:

[0012] (1) By introducing the joint input of multimodal sensor time series data from historical periods and weather forecast data from future periods, the time series prediction model (i.e., LSTM model) is used to accurately predict future photovoltaic power generation and power supply load demand. This can effectively cope with the volatility of solar energy resources and the uncertainty of user load demand, and can be used for pre-scheduling to optimize the output of the photovoltaic system in advance, avoiding frequent scheduling operations caused by the mismatch between load and power supply.

[0013] (2) Based on the core node, the data sent by each edge node is globally analyzed, the target scheduling parameters of each photovoltaic array are updated, and the scheduling parameter configuration is optimized globally by comprehensively analyzing the expected power generation and load demand of multiple photovoltaic arrays. In this way, not only can the power supply pressure between different photovoltaic arrays be balanced, but also the global performance degradation caused by local optimality can be effectively avoided, thus ensuring the overall power generation performance of the photovoltaic system.

[0014] (3) By combining multimodal sensor data with time series prediction models, the system can quickly adjust through node prediction and global scheduling mechanisms even in the face of extreme weather changes or sudden load demands, ensuring the stable operation of the photovoltaic system. It has stronger robustness and anti-interference capabilities, can effectively cope with changing external environments, and ensure the reliability of the system under complex working conditions.

[0015] Through this technical solution, edge network technology is used to achieve flexible and adjustable control of distributed photovoltaic systems. The core node is used to perform global analysis on the initial scheduling parameters of each edge node to generate target scheduling parameters, and the synergy between different arrays is comprehensively considered to achieve the global optimal scheduling strategy. Therefore, the implementation of a globally optimized collaborative control method in a large-scale distributed photovoltaic system can effectively reduce mutual interference between arrays and enhance the overall coordination and reliability of the distributed photovoltaic system. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, a brief introduction will be given below to the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0017] Figure 1 A flowchart showing an example of a flexible and adjustable control method for a distributed photovoltaic system according to an embodiment of the present application is shown;

[0018] Figure 2 An operation flow chart of an example of generating target scheduling parameters for each edge node based on a core node according to an embodiment of the present application is shown;

[0019] Figure 3 A schematic diagram of a structural connection of an example of a multi-scale attention LSTM model according to an embodiment of the present application is shown;

[0020] Figure 4 An operation flow chart showing an example of predicting initial scheduling parameters of a corresponding photovoltaic array according to a scheduling decision model according to an embodiment of the present application;

[0021] Figure 5 A structural block diagram of an example of a flexible adjustable control device for a distributed photovoltaic system according to an embodiment of the present application is shown;

[0022] Figure 6 It is a schematic structural diagram of an embodiment of an electronic device of the present application. DETAILED DESCRIPTION

[0023] In order to make the purpose, technical solution and advantages of the embodiments of the present application clearer, the technical solution in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.

[0024] In the technical solution of this application, the collection, storage, use, processing, transmission, provision and disclosure of user personal information involved shall comply with the provisions of relevant laws and regulations and shall not violate public order and good morals.

[0025] Figure 1 A flowchart of an example of a flexible and adjustable control method for a distributed photovoltaic system according to an embodiment of the present application is shown.

[0026] It should be understood that the edge computing network includes a core node and multiple edge nodes. Each edge node and core node has a certain computing capability. The core node can be used to manage each edge node. The edge node can share a certain resource processing pressure for the core node, and can support the implementation of some localized processing, and is faster and more efficient.

[0027] In some examples of the embodiments of the present application, a distributed photovoltaic system includes multiple photovoltaic arrays, and a corresponding sensor network and edge nodes are deployed for each photovoltaic array.

[0028] Regarding the executor of the method of the embodiment of the present application, it can be any controller or processor with computing or processing capabilities. By introducing a flexible and adjustable control method based on a timing prediction model, comprehensive utilization of multimodal data, dynamic timing prediction, global optimization scheduling decisions and regulation of edge computing networks, the adaptability, efficiency and stability of distributed photovoltaic systems are significantly improved, the rigidity problem of fixed parameter control strategies in the prior art is successfully solved, and efficient, flexible and intelligent management of photovoltaic systems is achieved.

[0029] In some examples, it can be an edge node, which is integrated and configured in an electronic device or terminal through software, hardware, or a combination of software and hardware, and the type of the terminal or electronic device can be diverse, such as a mobile phone, a tablet computer, or a desktop computer, etc.

[0030] like Figure 1 As shown, in step S110, multimodal sensing time series data of the deployed photovoltaic array in a historical preset time period and weather forecast data corresponding to a future preset time period are obtained.

[0031] Here, the time period length of the historical preset time period and the time period length of the future preset time period may be diverse and may be adjusted according to business requirements.

[0032] In some embodiments, a corresponding multimodal sensor network is deployed for each photovoltaic array to continuously collect key parameters in the photovoltaic array operating environment, such as solar radiation intensity, ambient temperature, power generation, grid operating parameters, and power supply load requirements. Then, the edge node performs preliminary filtering and denoising on the data collected by the sensor to ensure the accuracy and reliability of the data. For each sensor, the data is stored in a time series format to form multimodal sensing time series data, which contains multiple adjacent historical time steps and corresponding multimodal sensing data. In addition, in order to ensure the accuracy of the data, weather forecast data can be obtained based on a real-time update mechanism by interacting with the weather server, which mainly contains forecast information on future solar radiation intensity and ambient temperature.

[0033] In step S120, the multimodal sensing time series data and weather forecast data are input into a time series prediction model to predict the expected photovoltaic power generation and expected power supply load demand of the corresponding photovoltaic array in a preset time period in the future. The time series prediction model adopts an LSTM model.

[0034] Here, the LSTM model can be trained offline with a large amount of historical data and continuously optimized. When applied online, the LSTM model will gradually predict the photovoltaic power generation and load demand in the future preset time period based on the input multimodal time series data and weather forecast data. Since LSTM can effectively capture the long-term dependencies in the time series, it improves the prediction accuracy of photovoltaic power generation and power supply load demand and reduces the prediction error.

[0035] In step S130, the corresponding estimated photovoltaic power generation and the estimated power supply load demand are processed based on the scheduling decision model to predict the initial scheduling parameters of the corresponding photovoltaic array, and the parameter types of the scheduling parameters include the target output power and the grid-connected voltage adjustment value.

[0036] Here, the scheduling decision model can adopt various machine learning models, such as deep learning models or mathematical physics models, etc., with the predicted power generation and load demand as input, combined with the operating status of the photovoltaic array and the real-time parameters of the power grid, to calculate the initial scheduling parameters. Exemplarily, the scheduling decision model continuously optimizes the initial scheduling parameters based on historical scheduling results and current forecast data to ensure that the optimal output power and voltage regulation can be provided under different array environmental conditions. The optimal power generation utilization rate for the photovoltaic array is achieved through the output power setting value. The grid-connected voltage between the photovoltaic array and the power grid is adjusted through the grid-connected voltage adjustment value to ensure the stability of the power grid.

[0037] In step S140, the estimated photovoltaic power generation, the estimated power supply load demand and the initial scheduling parameters are sent to the core node, so that the core node performs a global analysis on the data sent by each edge node to generate a target scheduling parameter for each edge node.

[0038] In some embodiments, the edge nodes send their respective estimated photovoltaic power generation, power supply load demand and initial scheduling parameters to the core node through a dedicated communication network, so that the core node collects data from all edge nodes, uses a global scheduling algorithm to uniformly analyze the initial scheduling parameters of each photovoltaic array, and comprehensively considers the correlation between each photovoltaic array and the overall power grid demand to generate a globally optimized target scheduling parameter. Thus, the core node updates the target scheduling parameters of each photovoltaic array by analyzing the global data, making it more in line with actual needs, effectively avoiding the problem of local optimality, and improving the overall scheduling efficiency of the photovoltaic system.

[0039] In step S150, the corresponding target scheduling parameters are received from the core node, and the corresponding photovoltaic array is regulated in a future preset time period according to the target scheduling parameters.

[0040] In some implementations, after receiving the target scheduling parameters sent by the core node, the edge node applies them to the corresponding photovoltaic array control system to adjust the output power and grid-connected voltage of each photovoltaic array in real time. In the future preset time period, the system continuously adjusts the operating status of the photovoltaic array according to the target scheduling parameters to ensure that the output meets the grid demand and maximizes the use of solar energy resources.

[0041] Through the embodiments of the present application, by applying globally optimized target scheduling parameters, the photovoltaic array can achieve efficient operation under various working conditions, ensuring that the output power and grid-connected voltage meet the requirements of the power grid. Through the real-time regulation and feedback mechanism of the edge nodes, the photovoltaic system can dynamically respond to environmental changes and power grid demands, continuously optimize the operating status, and achieve the goal of flexible and adjustable control.

[0042] Therefore, by combining the advantages of edge computing and hierarchical decision-making, multi-level scheduling from local rapid response to global optimization is achieved, which can maintain the efficient operation of the system in a complex and changing environment and cope with the uncertainty of photovoltaic power generation.

[0043] Figure 2 An operational flowchart of an example of generating target scheduling parameters for each edge node based on a core node according to an embodiment of the present application is shown.

[0044] like Figure 2 As shown, in step S210, a photovoltaic scheduling diagram structure is constructed according to the estimated photovoltaic power generation, the estimated power supply load demand and the initial scheduling parameters sent by each edge node.

[0045] Here, the photovoltaic scheduling graph structure contains multiple graph nodes and edge connections. Each graph node uniquely corresponds to an edge node. The node characteristics of the graph node are defined by the corresponding expected photovoltaic power generation, expected power supply load demand and initial scheduling parameters. In addition, the edge weight of each edge connection is determined by the supply and demand complementarity, power transmission loss rate and historical coordination efficiency between the corresponding photovoltaic arrays of the connected graph nodes.

[0046] Here, the size of the edge weight can directly reflect the strength of the association between the corresponding photovoltaic arrays. The complementarity of power supply and demand between the two photovoltaic arrays is evaluated by the supply and demand complementarity. The higher the complementarity, the greater the edge weight. The loss rate in the power transmission process between the two arrays is calculated by the power transmission loss rate. The smaller the power transmission loss rate, the greater the edge weight. The historical collaborative efficiency can be calculated through historical data, which can evaluate the collaborative work efficiency between the two arrays. The higher the collaborative efficiency, the greater the edge weight.

[0047] By constructing a photovoltaic scheduling graph structure, the complex relationships of distributed photovoltaic systems are abstracted into a graph model. The multi-dimensional calculation of edge weights ensures the comprehensiveness of the relationships between photovoltaic arrays, covering key factors such as supply and demand matching, power transmission losses, and historical coordination. It can better model the relationships between photovoltaic arrays and explore potential collaborative optimization opportunities.

[0048] In step S220, the photovoltaic scheduling graph structure is processed based on the graph neural network to update the node characteristics of each graph node.

[0049] In some embodiments, advanced network architectures such as graph convolutional networks (GCN) or deep graph neural networks can be used. Through the graph neural network, the feature information of each node is transferred and fused with the feature information of its adjacent nodes through the edge connection of the graph. In each layer of the graph neural network, the node feature vector is weighted and summed with the feature vector of the neighboring node. After being processed by the multi-layer graph neural network, the initial node features will be updated to more expressive high-order features. The graph neural network uses the information propagation mechanism, which not only considers the information of a single node, but also integrates the information of neighboring nodes and the global graph structure, capturing the local and global information of each node, making the update of node features more comprehensive and accurate.

[0050] In step S230 , the node characteristics of each updated graph node are processed based on the global scheduling optimization model to determine the target scheduling parameters of each edge node.

[0051] Here, the global scheduling optimization model adopts a genetic algorithm model. The population individuals of the genetic algorithm model are defined by the scheduling parameter set of all photovoltaic arrays in the distributed photovoltaic system, and each gene in the population individual is defined by the scheduling parameter of the corresponding photovoltaic array. Specifically, the scheduling parameter set of all photovoltaic arrays in the distributed photovoltaic system is defined as a population individual, and each individual is composed of a series of genes. In the genetic algorithm, a gene represents a specific parameter in the scheduling scheme, such as the target output power and grid-connected voltage regulation value and other scheduling parameters. The initial population can be generated randomly or based on historical data.

[0052] In addition, the fitness function is defined based on the global supply and demand balance, global grid stability coefficient and global transmission loss rate of the distributed photovoltaic system. The supply and demand balance reflects the degree of matching between power generation and demand, the grid stability coefficient measures the operational stability of the entire power grid, and the transmission loss rate evaluates the loss of power transmission. Furthermore, the fitness of each population individual (i.e., the scheduling parameter set) is evaluated by the weighted sum of the above three. The higher the fitness, the more the scheduling parameter set of the individual can meet the global optimization requirements of the system.

[0053] Furthermore, the next generation of population is generated through selection, crossover and mutation operations. The selection operation gives priority to individuals with higher fitness according to fitness. The crossover operation generates new individuals by exchanging some genes of two parent individuals, and the mutation operation randomly changes certain genes to increase population diversity. Through multiple generations of iterations of the genetic algorithm, it gradually converges to the optimal solution, thereby determining the optimal individual and obtaining the target scheduling parameters of each photovoltaic array. Therefore, by using the global search and optimization of the genetic algorithm, the genetic algorithm simulates the process of natural selection to handle complex multi-objective optimization problems, and can search for scheduling parameters close to the global optimal in a huge solution space. The system can generate the optimal scheduling parameter combination to ensure that the photovoltaic system operates efficiently and in a complex power grid environment.

[0054] Through the embodiments of the present application, the scheduling problem of the photovoltaic system is modeled as a graph structure and features are extracted using a graph neural network, combined with a genetic algorithm for global optimization. By comprehensively considering the supply and demand balance, grid stability, and global transmission loss rate, the global scheduling optimization of the distributed photovoltaic system can be effectively achieved, ensuring the coordination between the various photovoltaic arrays, quickly determining the optimal scheduling parameters, and ensuring the stability and efficient operation of the system.

[0055] In some examples of the embodiments of the present application, the time series prediction model can be enhanced and designed in combination with the technical scenario. More specifically, the time series prediction model can adopt a multi-scale attention LSTM model. Here, the multi-scale attention LSTM model combines a multi-scale LSTM layer and a hybrid attention mechanism to improve the prediction accuracy of photovoltaic power generation and power supply load demand in distributed photovoltaic systems by processing data of different time scales and dynamically weighted input features.

[0056] Figure 3 A structural connection diagram of an example of a multi-scale attention LSTM model according to an embodiment of the present application is shown.

[0057] like Figure 3 As shown, the multi-scale attention LSTM model 300 includes a multi-scale LSTM layer 310, a hybrid attention mechanism layer 320 and an output layer 330.

[0058] The multi-scale LSTM layer 310 includes multiple LSTM layers (3111, 3113 ... 311n) and a fusion layer 312, each LSTM layer having a corresponding time scale. Here, the LSTM layer of each time scale is used to process input data of different time windows to generate corresponding scale hidden states.

[0059] Here, different laws of PV power generation and load demand in the short, medium and long term are captured through LSTM layers of different time scales. These time scales correspond to different time windows, such as the short time scale is hourly, the medium time scale is daily, and the long time scale is monthly or seasonal.

[0060] More specifically, the scaled hidden state of each time window is calculated as follows:

[0061]

[0062] In the formula, represents the scaled hidden state at the t-th time step under the time scale s, represents the multimodal sensor time series data at the historical time step t′, represents the weather forecast data at the future time step t″; represents the hidden state at the t-1th time step.

[0063] By introducing multi-scale LSTM layers, the model can process input data at different time scales, thereby finely capturing the different changing trends of photovoltaic power generation and power supply load demand in the short, medium and long term. For example, the short-time scale LSTM layer can effectively respond to sudden weather changes, while the long-time scale LSTM layer can identify seasonal changes and long-term trends, making the forecast more accurate and comprehensive.

[0064] The hybrid attention mechanism layer 320 includes a time attention module 321 and a feature attention module 323. The time attention module 321 dynamically adjusts the weight of each time step, and the feature attention module 323 performs weighting according to the importance of different features.

[0065] More specifically, the hybrid attention mechanism is expressed as follows:

[0066]

[0067]

[0068] In the formula, represents the temporal attention weight of the t′th historical time step under the time scale s, and They represent the weight matrix and bias term of the temporal attention mechanism corresponding to the time scale s; tanh represents the hyperbolic tangent function, represents the hidden state of the t′th historical time step, T H Represents the total number of historical time steps corresponding to the multimodal sensor time series data; represents the feature attention weight of the jth feature at time scale s; represents the feature representation of the jth feature in the historical multimodal time series data at all historical time steps, is the feature representation of the jth feature in the weather forecast data at all future time steps, yes and The joint feature representation of and They represent the weight matrix and bias term of the feature attention mechanism corresponding to the time scale s, and n represents the total number of features; represents the weighted hidden state at time scale s, Represents the hidden state corresponding to feature j at the historical time step t′ and time scale s.

[0069] The hybrid attention mechanism is used to weight time steps and input features to dynamically adjust the model's attention to important information. By applying the time attention mechanism, the weight of each time step is dynamically adjusted to ensure that the model can focus on the most critical time step when predicting. By applying the feature attention mechanism, weighting is performed according to the importance of different features. The model can dynamically adjust the attention to different features according to the actual situation and improve the prediction accuracy.

[0070] The output layer 330 is used to fuse the weighted hidden states of each scale to predict the expected photovoltaic power generation and the expected power supply load demand of the corresponding photovoltaic array in the future preset time period.

[0071] Here, a comprehensive representation is formed by integrating information from different time scales to predict future PV power generation and load demand.

[0072] More specifically, the structure of the output layer is expressed as follows:

[0073]

[0074] In the formula, is the fused multi-scale weighted hidden state, γ (s) represents the scale fusion weight of time scale s, and S represents the total number of time scales corresponding to the multi-scale LSTM layer; The estimated photovoltaic power generation and estimated power supply load demand are the final outputs of the model; W o and b o They represent the weight matrix and bias term of the output layer respectively.

[0075] Here, by introducing the time scale weight γ (s) , the user can make appropriate adjustments to the decision-making process of the model to control the impact of information at different time scales on the final prediction, making the model more flexible in practical applications and able to be adjusted according to specific needs.

[0076] In some examples of the embodiments of the present application, the multi-scale attention LSTM model uses mean square error (MSE) as a loss function to minimize the error between the predicted value and the true value. In addition, in the training optimization, an adaptive learning rate optimization algorithm (such as Adam) can be used for model training to accelerate convergence and avoid falling into local optimality.

[0077] Here, on the one hand, the scale fusion weight γ (s) These weights can be set by prior knowledge. For example, if it is known that a certain time scale (such as seasonal changes) is particularly important for prediction, this time scale can be artificially given a higher weight. On the other hand, it can also be calculated by calculating the time scale score of each time scale. The calculation of this time scale score can be based on the hidden state of each time scale, and then the score is converted into a weight γ by using the Softmax function. (s) , thus ensuring that the sum of the weights of all time scales is 1.

[0078] More preferably, in some examples of the embodiments of the present application, the scale fusion weight is obtained by calculating the following formula:

[0079]

[0080]

[0081] In the formula, g (s) Indicates the time scale score; A t represents the array state vector at time step t, which is determined according to the power generation fluctuation amplitude and load change rate at time step t; W γ and b γ Respectively represent the weight matrix and bias term related to the time scale, V γ is the array state vector A t The associated weight matrix; The index value representing the sum of the scores for all time scales.

[0082] In the embodiment of the present application, the array state vector A t Taking into account the fluctuation amplitude of power generation and the load change rate, the actual operating status of the array at different time steps can be more accurately reflected. The dynamic state representation enables the model to respond to the actual needs and changes of the system in a timely manner during the prediction, thereby improving the real-time and accuracy of the prediction. For example, when the system load suddenly increases or the power generation fluctuates drastically, the array state vector can guide the model to pay more attention to changes in short time scales, ensuring that the scheduling strategy can quickly respond to changing environmental conditions and maintain the stability of the system. Therefore, the array state vector dynamically adjusts the γ in the multi-scale LSTM model. (s), which enables the model to flexibly distribute weights between different time scales and improves the effect of multi-scale information fusion.

[0083] It should be noted that the relationship between photovoltaic power generation, power supply load demand and scheduling parameters is usually nonlinear, and it may be difficult to accurately capture system behavior using a linear model directly. By segmenting the nonlinear relationship and performing linear approximation in each interval, the prediction accuracy can be improved while maintaining computational efficiency.

[0084] In view of this, in some examples of the embodiments of the present application, for the scheduling decision model, it includes multiple local linear model modules, each of which has a corresponding supply and demand parameter interval. In a distributed photovoltaic system, the division of the supply and demand parameter interval is to capture the nonlinear relationship between photovoltaic power generation and power supply load demand. By dividing these parameters into multiple intervals, a local linear model can be used in each interval for approximate processing, thereby improving the accuracy and adaptability of scheduling decisions.

[0085] Figure 4 An operational flow chart of an example of predicting initial scheduling parameters of a corresponding photovoltaic array according to a scheduling decision model according to an embodiment of the present application is shown.

[0086] like Figure 4 As shown, in step S410, a target supply and demand parameter interval matching the expected photovoltaic power generation and the expected power supply load demand is determined from multiple supply and demand parameter intervals.

[0087] Specifically, a plurality of supply and demand matrix intervals are divided according to photovoltaic power generation and power supply load demand, and each supply and demand parameter matrix interval corresponds to a local linear model to describe the approximate linear relationship within the interval.

[0088] In some embodiments, by using historical data to perform statistical analysis on photovoltaic power generation and power supply load demand, the parameter range is divided into several intervals according to the distribution characteristics of the data, such as equal-interval division, equal-frequency division, or division based on data clustering.

[0089] In step S420, the estimated photovoltaic power generation and the estimated power supply load demand are input into a target local linear model module that matches the target supply and demand parameter interval to determine a first scheduling parameter.

[0090] Through numerical matching, the matching local linear model module can be found quickly. Through segmented processing, the model can simplify the calculation in a linear manner in each interval, and at the same time, it can better approximate the nonlinear relationship as a whole.

[0091] It should be noted that the training of each local linear model module can be carried out by using historical data through least squares method or other regression methods to obtain the linear weights and biases of the model modules in each interval. In addition, during the operation of the model, the weights and biases of each interval can be dynamically adjusted based on the actual feedback of the system to ensure that the model can continue to adapt to changes in the actual situation.

[0092] In step S430, it is detected whether the first scheduling parameter meets the preset array safety operation condition.

[0093] In step S441, if the first scheduling parameter satisfies the array safe operation condition, an initial scheduling parameter is determined according to the first scheduling parameter.

[0094] In step S443, if the first scheduling parameter does not meet the array safety operation condition, the first scheduling parameter is updated according to the array safety operation condition to determine the initial scheduling parameter.

[0095] Here, the array safe operation conditions include the grid voltage safety range and the power limit safety range. Exemplarily, if the grid voltage or output power exceeds the safety range, the voltage or output power is adjusted accordingly to the safety range. Through the grid voltage safety range, the grid voltage of the array must be kept within the safety range to avoid being too high or too low. Through the power limit safety range, the output power should be within the rated power range of the array equipment to avoid overload operation. Therefore, through the fusion rule-driven mechanism, the model can detect whether the preset array safety operation conditions are met after determining the initial scheduling parameters, thereby ensuring the safety and stability of the system under various abnormal conditions.

[0096] Through the embodiments of the present application, by dividing the photovoltaic power generation and power supply load demand into multiple supply and demand parameter intervals, and using independent local linear models in each interval, the model can better capture the nonlinear behavior of the photovoltaic system, and the accuracy of the scheduling decision can be effectively improved based on the segmented processing method. In addition, the piecewise linearization model can make corresponding scheduling decisions for different supply and demand conditions, especially in scenarios where the photovoltaic power generation and load demand fluctuate greatly. The model can adaptively select appropriate local linear models for calculation, thereby ensuring that the system can achieve optimal scheduling under various operating conditions. In addition, due to the combination of piecewise linearization and rule-driven, the design of the scheduling decision model has lightweight characteristics, which is particularly suitable for deployment on edge nodes. The model can run efficiently with limited computing resources, reduce the consumption of system resources, and can quickly complete pre-scheduling calculations under low power consumption conditions.

[0097] It should be noted that in a distributed photovoltaic system, the correlation and difference between the status and operating environment of each photovoltaic array are large, which means that each photovoltaic array has a different impact on the overall scheduling of the system. In order to capture these differences and achieve better global scheduling, the Graph Attention Network (GAT) is introduced, which enables the network to dynamically adjust the node characteristics of the node according to the importance of each photovoltaic array.

[0098] In a distributed photovoltaic system, each photovoltaic array has different contributions and influences on the global scheduling of the system. Through the graph neural network based on the attention mechanism, the system can automatically identify the key nodes that have a greater impact on the global scheduling and give these nodes a higher attention weight. This mechanism ensures that the characteristics of the key nodes can have a greater impact on the global scheduling during the update process, thereby optimizing the operating efficiency of the entire photovoltaic system.

[0099] In some examples of the embodiments of the present application, the graph neural network based on the attention mechanism updates the node features of the graph nodes in the following manner:

[0100]

[0101] In the formula, is the attention coefficient calculated for the mth attention head, a T represents the transpose of the weight vector used to calculate the attention coefficient, LeakyReLU represents the leakyRelu activation function, and W m represents the linear transformation matrix of the mth attention head; l u and l v denote the input feature vectors of graph nodes u and v respectively, ‖ denotes the connection operator, and μ denotes the adjustment coefficient; e uv represents the edge weight between graph node u and neighbor graph node v, Represents the set of neighbor nodes of graph node u; l u' represents the updated node feature vector of graph node u, M represents the total number of attention heads, and σ represents the Sigmoid activation function; It is a normalization factor, which represents the cumulative effect of all neighbor nodes k of node u.

[0102] In view of the explanation of the above formula, nonlinearity is introduced through the LeakyReLU activation function to prevent overly smooth feature representation. In order to enhance the expressiveness of the model, a multi-head attention mechanism is introduced in GAT, that is, multiple independent attention coefficients are calculated in parallel in different attention heads, and then the outputs of these attention heads are averaged, thereby improving the robustness and feature extraction ability of the model while keeping the output dimension unchanged. In addition, the edge weights of different edge connections are comprehensively considered in the process of node update. By introducing edge weights and multi-head attention mechanisms, the relationship strength between nodes can be more accurately reflected, so that the node update process can more accurately reflect the complex relationship between nodes.

[0103] By combining the attention mechanism with edge weights, graph neural networks can more carefully express the complex relationships between photovoltaic arrays during the node feature update process, achieve refined feature expression, and help the system better coordinate the scheduling decisions of various photovoltaic arrays on a global scale, ensuring the supply and demand balance of the system and the stability of the power grid.

[0104] In some examples of the embodiments of the present application, the edge weight is calculated by the following formula:

[0105]

[0106] In the formula, C uv is the supply-demand complementarity, which indicates the complementarity of the photovoltaic arrays corresponding to the graph nodes u and v in terms of power generation and load demand; η uv is the power transmission loss rate, which indicates the loss rate when power is transmitted from the photovoltaic array corresponding to the node u to the photovoltaic array corresponding to the node v; B uv is the historical coordination efficiency, which indicates the efficiency of the historical coordination of the photovoltaic arrays corresponding to the graph node u and the graph node v; 1 The first adjustment parameter, λ, represents the relative importance of balancing supply-demand complementarity and historical synergy efficiency. 2 represents the second adjustment parameter used to adjust the influence of transmission loss rate in edge weight; J pred,u and J pred,v They represent the estimated power generation of the photovoltaic array corresponding to the graph node u and the graph node v, respectively. pred,u and L pred,v They represent the estimated power supply load demand corresponding to the graph node u and the graph node v respectively; δ uv It represents the line transmission efficiency coefficient between the photovoltaic arrays corresponding to the graph node u and the graph node v, d uv represents the transmission line distance between the photovoltaic arrays corresponding to the graph node u and the graph node v; ξ uv (t) represents the matching degree of output power when the photovoltaic arrays corresponding to the graph node u and the graph node v work together at the historical time step t; TD Represents the total number of historical time steps of the PV array working together.

[0107] In view of the explanation of the above formula (11), the multiplication structure of supply and demand complementarity, transmission efficiency and historical synergy efficiency can better reflect the synergy of these factors. If any factor is zero, the edge weight will be significantly reduced. uv The linear weighted term of is added to make the negative impact of the transmission loss rate on the edge weight more prominent, especially when the transmission loss rate is high. This can more reasonably reduce the weight of the high-loss path.

[0108] By designing the edge weight calculation formula e uv ,Taking supply-demand complementarity, power transmission loss and historical coordination into ,comprehensive consideration, the calculation of edge weights can provide more refined ,node feature update information, which enables the graph neural network to perform ,better in global scheduling optimization, and more accurately reflects the ,complex relationship between photovoltaic arrays in the photovoltaic scheduling graph structure, which helps ,improve the overall operation efficiency and safety of the system.

[0109] Specifically, by introducing the complementarity of supply and demand, the system can give priority to those node pairs with good supply and demand matching, which helps to establish more effective power dispatching channels between photovoltaic arrays, optimize the system's balancing ability, and reduce the imbalance of supply and demand in power dispatching. Through the design of the power transmission loss rate part, it is ensured that the system can give priority to paths with smaller transmission losses during dispatching, thereby reducing energy waste, improving the energy utilization efficiency of the entire distributed photovoltaic system, and thus improving the economy and environmental benefits of the system. Through the historical collaborative efficiency part, the system can consider historically successful collaborative working modes in dispatching decisions. The node combination that has successfully coordinated historically will obtain a higher weight in the edge weight calculation, thereby enhancing the reliability and stability of the dispatching strategy, enabling the system to better cope with complex and changing operating environments and ensure the efficient execution of dispatching strategies.

[0110] In some examples of the embodiments of the present application, by designing the fitness function of the genetic algorithm model, the overall scheduling efficiency and operation stability of the system can be effectively improved by comprehensively considering the supply and demand balance, grid stability and transmission loss.

[0111] More specifically, the fitness function F of the global scheduling optimization model is fitness It is expressed by the following formula:

[0112] F fitness =ψ 1 ·G S +ψ 2 ·G G +ψ 3 ·G T , formula (15)

[0113]

[0114] In the formula, G S , G G and G T They represent the global supply-demand balance, global power grid stability coefficient and global transmission loss rate, respectively, 1 , 2 and ψ 3 They represent the corresponding importance adjustment weights, ∈ represents the preset constant; P out,u is the output power of the photovoltaic array corresponding to the u-th graph node after update, L demand,u is the power supply load demand of the photovoltaic array corresponding to the u-th graph node, Z represents the total number of photovoltaic arrays in the distributed photovoltaic system; V grid,u is the grid-connected voltage of the photovoltaic array corresponding to the uth graph node, V nominal Indicates the nominal voltage of the power grid.

[0115] Regarding the explanation of formula (16), represents the total output power of all PV arrays in the system, and The total power demand of the system is expressed by Normalize the difference so that the supply and demand balance can be reflected as a standardized value, so that G S It can measure the degree of matching between the total power generation of the system and the total demand. When the supply and demand are completely balanced (that is, the total output power is equal to the total demand), G S Taking 1 means that the system has achieved the best supply and demand balance; when supply and demand are unbalanced, G S The value will decrease.

[0116] Regarding the explanation of formula (17), Measures the deviation between the actual grid-connected voltage of the u-th photovoltaic array and the nominal voltage. By summing the deviations of all photovoltaic arrays, G G Reflects the overall voltage stability of the system. G It is a reverse indicator. The larger the deviation in the denominator, the higher the G G The smaller the value, the worse the grid stability. G The higher it is, the more stable the grid operation is.

[0117] Regarding the explanation of formula (18), η uv ·e uv It represents the weighted product of the transmission loss between nodes and the edge weight. By accumulating all possible node pairs, the total transmission loss of the entire system is calculated. G Similar to G. T It is a reverse indicator. The larger the weighted sum in the denominator, the higher the GT The smaller the value of, the higher the global transmission loss rate.

[0118] Through the fitness function in the embodiment of the present application, based on the global supply and demand balance G S , the model can continuously adjust the output power of each photovoltaic array to maximize the supply and demand balance of the entire system. G The model can optimize the grid-connected voltage of each photovoltaic array, ensure the stability of the system voltage in each area, and prevent the impact of voltage fluctuations on system operation. T ,The model can give priority to the path with low transmission loss in the ,global scheduling process, reduce energy waste, and improve the overall ,efficiency of the system.,Therefore, the genetic algorithm continuously optimizes the population individuals ,through evolutionary operations, and through comprehensive consideration of ,supply demand balance, grid stability and transmission loss, the scheduling parameters can adapt to ,variable system conditions, which can effectively improve the overall scheduling ,efficiency and operation stability of the system.

[0119] It should be noted that, for the aforementioned method embodiments, for the sake of simplicity of description, they are all expressed as a series of actions combined, but those skilled in the art should be aware that the present application is not limited by the described order of actions, because according to the present application, certain steps can be performed in other orders or simultaneously. Secondly, those skilled in the art should also be aware that the embodiments described in the specification are all preferred embodiments, and the actions and modules involved are not necessarily required by the present application. In the above embodiments, the description of each embodiment has its own emphasis. For parts that are not described in detail in a certain embodiment, please refer to the relevant description of other embodiments.

[0120] Figure 5 A structural block diagram of an example of a flexible and adjustable control device for a distributed photovoltaic system according to an embodiment of the present application is shown.

[0121] The distributed photovoltaic system includes multiple photovoltaic arrays, and a corresponding sensor network and edge node are deployed for each photovoltaic array. The flexible adjustable control device is set at the edge node.

[0122] like Figure 5 As shown, the flexible adjustable control device 500 of the distributed photovoltaic system includes an acquisition unit 510, a supply and demand prediction unit 520, an initial scheduling prediction unit 530, a data uploading unit 540 and a target scheduling control unit 550.

[0123] The acquisition unit 510 is used to acquire the multimodal sensing time series data of the deployed photovoltaic array in a historical preset time period and the weather forecast data corresponding to the future preset time period; the multimodal sensing time series data includes multiple adjacent historical time steps and corresponding multimodal sensing data, and the parameter types of the multimodal sensing data include: solar radiation intensity, ambient temperature, power generation, power grid operation parameters and power supply load requirements.

[0124] The supply and demand prediction unit 520 is used to input the multimodal sensing time series data and the weather forecast data into a time series prediction model to predict the expected photovoltaic power generation and expected power supply load demand of the corresponding photovoltaic array in the future preset time period; the time series prediction model adopts an LSTM model.

[0125] The initial scheduling prediction unit 530 is used to process the corresponding expected photovoltaic power generation and the expected power supply load demand based on the scheduling decision model to predict the initial scheduling parameters of the corresponding photovoltaic array; the parameter types of the scheduling parameters include the target output power and the grid-connected voltage adjustment value.

[0126] The data upload unit 540 is used to send the estimated photovoltaic power generation, the estimated power supply load demand and the initial scheduling parameters to the core node, so that the core node performs a global analysis on the data sent by each edge node to generate a target scheduling parameter for each edge node.

[0127] The target scheduling control unit 550 is used to receive the corresponding target scheduling parameters from the core node, and regulate the corresponding photovoltaic array in the future preset time period according to the target scheduling parameters.

[0128] In some embodiments, an embodiment of the present application provides a non-volatile computer-readable storage medium, which stores one or more programs including execution instructions, and the execution instructions can be read and executed by an electronic device (including but not limited to a computer, a server, or a network device, etc.) to execute the steps of any of the above-mentioned methods for flexible and adjustable control of a distributed photovoltaic system in the present application.

[0129] In some embodiments, the embodiments of the present application also provide a computer program product, which includes a computer program stored on a non-volatile computer-readable storage medium, and the computer program includes program instructions. When the program instructions are executed by a computer, the computer executes the steps of any one of the above-mentioned flexible and adjustable control methods for distributed photovoltaic systems.

[0130] In some embodiments, an embodiment of the present application also provides an electronic device, comprising: at least one processor, and a memory communicatively connected to the at least one processor, wherein the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to perform the steps of the flexible and adjustable control method of a distributed photovoltaic system.

[0131] Figure 6 is a schematic diagram of the hardware structure of an electronic device for executing a flexible and adjustable control method for a distributed photovoltaic system provided by another embodiment of the present application, such as Figure 6 As shown, the device includes:

[0132] One or more processors 610 and memory 620, Figure 6 A processor 610 is taken as an example.

[0133] The device for executing the flexible and adjustable control method for a distributed photovoltaic system may further include: an input device 630 and an output device 640 .

[0134] The processor 610, the memory 620, the input device 630 and the output device 640 may be connected via a bus or other means. Figure 6 The example of connecting through bus is taken in the following.

[0135] The memory 620, as a non-volatile computer-readable storage medium, can be used to store non-volatile software programs, non-volatile computer executable programs and modules, such as program instructions / modules corresponding to the flexible and adjustable control method of the distributed photovoltaic system in the embodiment of the present application. The processor 610 executes various functional applications and data processing of the server by running the non-volatile software programs, instructions and modules stored in the memory 620, that is, realizing the flexible and adjustable control method of the distributed photovoltaic system in the above method embodiment.

[0136] The memory 620 may include a program storage area and a data storage area, wherein the program storage area may store an operating system, an application required for at least one function; the data storage area may store data created according to the use of the electronic device, etc. In addition, the memory 620 may include a high-speed random access memory, and may also include a non-volatile memory, such as at least one disk storage device, a flash memory device, or other non-volatile solid-state storage device. In some embodiments, the memory 620 may optionally include a memory remotely arranged relative to the processor 610, and these remote memories may be connected to the electronic device via a network. Examples of the above-mentioned network include, but are not limited to, the Internet, an intranet, a local area network, a mobile communication network, and combinations thereof.

[0137] The input device 630 can receive input digital or character information and generate signals related to user settings and function control of the electronic device. The output device 640 can include a display device such as a display screen.

[0138] The one or more modules are stored in the memory 620 , and when executed by the one or more processors 610 , the flexible and adjustable control method of the distributed photovoltaic system in any of the above method embodiments is executed.

[0139] The above-mentioned product can execute the method provided in the embodiment of the present application, and has the functional modules and beneficial effects corresponding to the execution method. For technical details not fully described in this embodiment, please refer to the method provided in the embodiment of the present application.

[0140] The electronic devices of the embodiments of the present application exist in various forms, including but not limited to:

[0141] (1) Mobile communication equipment: This type of equipment is characterized by having mobile communication functions and its main purpose is to provide voice and data communications. This type of terminal includes: smart phones, multimedia phones, functional phones, and low-end phones.

[0142] (2) Ultra-mobile personal computer devices: These devices belong to the category of personal computers, have computing and processing functions, and generally also have mobile Internet access features. These terminals include: PDA, MID and UMPC devices, etc.

[0143] (3) Portable entertainment devices: These devices can display and play multimedia content. They include audio and video players, handheld game consoles, e-books, smart toys, and portable car navigation devices.

[0144] (4) Other onboard electronic devices with data interaction functions, such as on-board devices installed in vehicles.

[0145] The device embodiments described above are merely illustrative, wherein the units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0146] Through the description of the above implementation methods, those skilled in the art can clearly understand that each implementation method can be implemented by means of software plus a general hardware platform, and of course, by hardware. Based on this understanding, the above technical solution is essentially or the part that contributes to the relevant technology can be embodied in the form of a software product, and the computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, a disk, an optical disk, etc., including a number of instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods described in each embodiment or some parts of the embodiments.

[0147] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present application, rather than to limit it. Although the present application has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present application.

Claims

1. A flexible and adjustable control method for a distributed photovoltaic system, wherein the distributed photovoltaic system comprises a plurality of photovoltaic arrays, and a corresponding sensor network and an edge node are deployed for each photovoltaic array, the method comprising: Each edge node is used to perform the following operations: Acquire multimodal sensing time series data of the deployed photovoltaic array in a historical preset time period and weather forecast data corresponding to a future preset time period; the multimodal sensing time series data includes a plurality of adjacent historical time steps and corresponding multimodal sensing data, and the parameter types of the multimodal sensing data include: solar radiation intensity, ambient temperature, power generation, power grid operation parameters and power supply load demand; Inputting the multimodal sensing time series data and the weather forecast data into a time series prediction model to predict the expected photovoltaic power generation and the expected power supply load demand of the corresponding photovoltaic array in the future preset time period; the time series prediction model adopts an LSTM model; Based on the scheduling decision model, the corresponding expected photovoltaic power generation and the expected power supply load demand are processed to predict the initial scheduling parameters of the corresponding photovoltaic array; the parameter types of the scheduling parameters include the target output power and the grid-connected voltage adjustment value; Sending the estimated photovoltaic power generation, the estimated power supply load demand and the initial scheduling parameters to the core node, so that the core node performs a global analysis on the data sent by each edge node to generate a target scheduling parameter for each edge node; The corresponding target scheduling parameters are received from the core node, and the corresponding photovoltaic array is regulated in the future preset time period according to the target scheduling parameters.

2. The method according to claim 1, wherein: The core node is used to generate target scheduling parameters for each edge node by performing the following operations: A photovoltaic scheduling graph structure is constructed according to the estimated photovoltaic power generation, estimated power supply load demand and initial scheduling parameters sent by each edge node; the photovoltaic scheduling graph structure includes multiple graph nodes and edge connections; each of the graph nodes uniquely corresponds to an edge node, and the node characteristics of the graph node are defined by the corresponding estimated photovoltaic power generation, estimated power supply load demand and initial scheduling parameters; the edge weight of each edge connection is determined by the supply and demand complementarity, power transmission loss rate and historical coordination efficiency between the corresponding photovoltaic arrays of the connected graph nodes; Processing the photovoltaic scheduling graph structure based on a graph neural network to update node features of each of the graph nodes; Based on the global scheduling optimization model, the node characteristics of each updated graph node are processed to determine the target scheduling parameters of each edge node; the global scheduling optimization model adopts a genetic algorithm model; the population individuals of the genetic algorithm model are defined by the scheduling parameter set of all photovoltaic arrays in the distributed photovoltaic system, each gene in the population individual is defined by the scheduling parameter of the corresponding photovoltaic array, and the fitness function is defined according to the global supply and demand balance of the distributed photovoltaic system, the global power grid stability coefficient and the global transmission loss rate.

3. The method according to claim 2, wherein: The time series prediction model adopts a multi-scale attention LSTM model, which includes a multi-scale LSTM layer, a hybrid attention mechanism layer and an output layer; The multi-scale LSTM layer includes multiple LSTM layers and a fusion layer, each of which has a corresponding time scale; the LSTM layer of each time scale is used to process input data of different time windows to generate corresponding scale hidden states: In the formula, represents the scaled hidden state at the t-th time step under the time scale s, represents the multimodal sensor time series data at the historical time step t′, represents the weather forecast data at the future time step t″; represents the hidden state at the t-1th time step; The hybrid attention mechanism layer includes a time attention module and a feature attention module. The time attention module dynamically adjusts the weight of each time step, and the feature attention module performs weighting according to the importance of different features: In the formula, represents the temporal attention weight of the t′th historical time step under the time scale s, and They represent the weight matrix and bias term of the temporal attention mechanism corresponding to the time scale s; tanh represents the hyperbolic tangent function, represents the hidden state of the t′th historical time step, T H Represents the total number of historical time steps corresponding to the multimodal sensor time series data; represents the feature attention weight of the jth feature at time scale s; represents the feature representation of the jth feature in the historical multimodal time series data at all historical time steps, is the feature representation of the jth feature in the weather forecast data at all future time steps, yes and The joint feature representation of and They represent the weight matrix and bias term of the feature attention mechanism corresponding to the time scale s, and n represents the total number of features; represents the weighted hidden state at time scale s, represents the hidden state corresponding to feature j at the historical time step t′ and time scale s; The output layer is used to fuse the weighted hidden states of each scale to predict the expected photovoltaic power generation and the expected power supply load demand of the corresponding photovoltaic array in the future preset time period: In the formula, is the fused multi-scale weighted hidden state, γ (s) represents the scale fusion weight of time scale s, and S represents the total number of time scales corresponding to the multi-scale LSTM layer; The estimated photovoltaic power generation and estimated power supply load demand are the final outputs of the model; W o and b o Represent the weight matrix and bias term of the output layer respectively.

4. The method according to claim 3, wherein: The scale fusion weight is calculated by the following formula: In the formula, g (s) Indicates the time scale score; A t W represents the array state vector at time step t, and the array state vector is determined according to the power generation fluctuation amplitude and load change rate at time step t; W γ and b γ Respectively represent the weight matrix and bias term related to the time scale, V γ is the array state vector A t The associated weight matrix; The index value representing the sum of the scores for all time scales.

5. The method according to claim 2, wherein: The scheduling decision model includes a plurality of local linear model modules, each of which has a corresponding supply and demand parameter interval; the scheduling decision model is used to predict the initial scheduling parameters of the corresponding photovoltaic array by performing the following operations: Determining a target supply and demand parameter interval that matches the estimated photovoltaic power generation and the estimated power supply load demand from a plurality of supply and demand parameter intervals; Inputting the estimated photovoltaic power generation and the estimated power supply load demand into a target local linear model module that matches the target supply and demand parameter interval to determine a first scheduling parameter; Detecting whether the first scheduling parameter meets a preset array safety operation condition; the array safety operation condition includes a grid-connected voltage safety range and a power limit safety range; If the first scheduling parameters satisfy the array safe operation conditions, determining initial scheduling parameters according to the first scheduling parameters; as well as If the first scheduling parameter does not meet the array safe operation condition, the first scheduling parameter is updated according to the array safe operation condition to determine the initial scheduling parameter.

6. The method according to claim 2, wherein: The graph neural network adopts a graph neural network based on an attention mechanism, which is used to update the node features of the graph nodes in the following way: In the formula, is the attention coefficient calculated for the mth attention head, a T represents the transpose of the weight vector used to calculate the attention coefficient, LeakyReLU represents the leakyRelu activation function, and W m represents the linear transformation matrix of the mth attention head; l u and l v denote the input feature vectors of graph nodes u and v respectively, ‖ denotes the connection operator, and μ denotes the adjustment coefficient; e uv represents the edge weight between graph node u and neighbor graph node v, Represents the set of neighbor nodes of graph node u; l u' represents the updated node feature vector of graph node u, M represents the total number of attention heads, and σ represents the Sigmoid activation function; It is a normalization factor, which represents the cumulative effect of all neighbor nodes k of node u.

7. The method according to claim 6, wherein: The edge weight is calculated as follows: In the formula, C uv is the supply-demand complementarity, which indicates the complementarity of the photovoltaic arrays corresponding to the graph nodes u and v in terms of power generation and load demand; η uv is the power transmission loss rate, which indicates the loss rate when power is transmitted from the photovoltaic array corresponding to the node u to the photovoltaic array corresponding to the node v; B uv is the historical coordination efficiency, which indicates the efficiency of the historical coordination of the photovoltaic arrays corresponding to the graph nodes u and v; λ1 represents the first adjustment parameter used to balance the relative importance of the supply-demand complementarity and the historical coordination efficiency, and λ2 represents the second adjustment parameter used to adjust the influence of the transmission loss rate in the edge weight; J pred,u and J pred,v They represent the estimated power generation of the photovoltaic array corresponding to the graph node u and the graph node v, respectively. pred,u and L pred,v They represent the estimated power supply load demand corresponding to the graph node u and the graph node v respectively; δ uv It represents the line transmission efficiency coefficient between the photovoltaic arrays corresponding to the graph node u and the graph node v, d uv represents the transmission line distance between the photovoltaic arrays corresponding to the graph node u and the graph node v; ξ uv (t) represents the matching degree of output power when the photovoltaic arrays corresponding to the graph node u and the graph node v work together at the historical time step t; T D Represents the total number of historical time steps of the PV array working together.

8. The method according to claim 7, wherein: The fitness function F of the global scheduling optimization model is fitness It is expressed by the following formula: F fitness =ψ1·G S +ψ2·G G +ψ3·G T , In the formula, G S , G G and G T They represent the global supply-demand balance, the global power grid stability coefficient and the global transmission loss rate respectively, ψ1, ψ2 and ψ3 represent the corresponding importance adjustment weights respectively, ∈ represents the preset constant; P out,u is the output power of the photovoltaic array corresponding to the u-th graph node after update, L demand,u is the power supply load demand of the photovoltaic array corresponding to the u-th graph node, Z represents the total number of photovoltaic arrays in the distributed photovoltaic system; V grid,u is the grid-connected voltage of the photovoltaic array corresponding to the uth graph node, V nominal Indicates the nominal voltage of the power grid.

9. A flexible and adjustable control device for a distributed photovoltaic system, the distributed photovoltaic system comprising a plurality of photovoltaic arrays, and a corresponding sensor network and edge node are deployed for each photovoltaic array, the device comprising an acquisition unit, a supply and demand prediction unit, an initial scheduling prediction unit, a data uploading unit and a target scheduling control unit arranged in the edge node; An acquisition unit is used to acquire multimodal sensing time series data of the deployed photovoltaic array in a historical preset time period and weather forecast data corresponding to a future preset time period; the multimodal sensing time series data includes a plurality of adjacent historical time steps and corresponding multimodal sensing data, and the parameter types of the multimodal sensing data include: solar radiation intensity, ambient temperature, power generation, power grid operation parameters and power supply load demand; A supply and demand prediction unit, used for inputting the multimodal sensing time series data and the weather forecast data into a time series prediction model to predict the expected photovoltaic power generation and the expected power supply load demand of the corresponding photovoltaic array in the future preset time period; the time series prediction model adopts an LSTM model; The initial scheduling prediction unit is used to process the corresponding expected photovoltaic power generation and the expected power supply load demand based on the scheduling decision model to predict the initial scheduling parameters of the corresponding photovoltaic array; the parameter types of the scheduling parameters include the target output power and the grid-connected voltage adjustment value; A data uploading unit, used for sending the estimated photovoltaic power generation, the estimated power supply load demand and the initial scheduling parameters to the core node, so that the core node performs a global analysis on the data sent by each edge node to generate a target scheduling parameter for each edge node; The target scheduling control unit is used to receive the corresponding target scheduling parameters from the core node, and regulate the corresponding photovoltaic array in the future preset time period according to the target scheduling parameters.

Images