Power distribution network centralized cooperative control method, device and equipment based on photovoltaic power generation and reactive compensation and medium

Abstract

The invention discloses a power distribution network centralized cooperative control method and device based on photovoltaic power generation and reactive compensation, equipment and a medium, and relates to the technical field of cooperative control, and the method comprises the following steps: obtaining historical photovoltaic power generation data and environmental data of each reactive compensation node of a power distribution network, and carrying out the preprocessing; inputting the historical photovoltaic power generation data and the environmental data of each reactive compensation node of the power distribution network into a preset bidirectional LSTM model based on offset adjustment to obtain a power factor prediction value of each reactive compensation node; inputting the power factor prediction value of each reactive compensation node into a preset reactive power adjustment calculation formula to obtain a reactive power error value and a reactive power adjustment value; and grading the reactive power error value and correcting the reactive power of each reactive compensation node based on the reactive power adjustment value.

Description

Technical Field

[0001] The present invention relates to the field of collaborative control technology, and more specifically, to a centralized collaborative control method, device, equipment and medium for a distribution network based on photovoltaic power generation and reactive power compensation. Background Art

[0002] Under the current global energy transformation background, photovoltaic power generation, as a clean and renewable energy, has been widely used and developed rapidly. Especially under the policy of vigorously promoting the use of new energy, the penetration rate of photovoltaic power generation in the power system has been continuously improved, and more and more large-scale photovoltaic power stations and distributed photovoltaic systems have been connected to the distribution network.

[0003] The reactive power compensation nodes in the distribution network are interrelated. Individually controlling each node may lead to waste of resources or inefficiency, poor overall scheduling, and traditional distributed control methods are difficult to perceive the global state in real time, which may lead to problems such as local fault spread, frequency and voltage fluctuations. At the same time, the output power of the photovoltaic power generation system is mainly active power, but due to equipment characteristics and grid conditions, reactive power fluctuations may occur, which is easily affected by environmental factors such as clouds, temperature, humidity, etc., resulting in deviations and affecting the overall efficiency of the grid. In view of the above problems, the present invention proposes a solution. Summary of the invention

[0004] In order to overcome the above-mentioned defects of the prior art, the embodiments of the present invention provide a centralized collaborative control method, device, equipment and medium for a distribution network based on photovoltaic power generation and reactive power compensation, so as to solve the problem of affecting the overall efficiency of the distribution network due to individual control nodes and failure to consider environmental factors.

[0005] To achieve the above object, the present invention provides the following technical solutions: A centralized collaborative control method for a distribution network based on photovoltaic power generation and reactive power compensation comprises: obtaining historical photovoltaic power generation data and environmental data of each reactive power compensation node of the distribution network and performing preprocessing, wherein the historical photovoltaic power generation data of each reactive power compensation node of the distribution network comprises voltage data, current data, power data, and reactive power compensation node status data; inputting the historical photovoltaic power generation data and environmental data of each reactive power compensation node of the distribution network into a preset bidirectional LSTM model based on offset adjustment to obtain a power factor prediction value of each reactive power compensation node; inputting the power factor prediction value of each reactive power compensation node into a preset reactive power adjustment calculation formula to obtain a reactive power error value and a reactive power adjustment value; and grading the reactive power error value and correcting the reactive power of each reactive power compensation node based on the reactive power adjustment value.

[0006] In a preferred embodiment, the preset bidirectional LSTM model based on offset adjustment is specifically as follows: constructing a convolution layer, performing feature extraction on the historical photovoltaic power generation data and environmental data of each reactive compensation node of the distribution network, and obtaining a first power characteristic graph; based on the first power characteristic graph, using the correlation matrix method to calculate the offset, and obtaining the offset of the first power characteristic graph; inputting the offset of the first power characteristic graph and the first power characteristic graph into a preset variability convolution module for feature extraction, and obtaining a second power characteristic graph; using the second power characteristic graph as the input of the preset bidirectional LSTM model, and obtaining a bidirectional LSTM model based on offset adjustment.

[0007] In a preferred embodiment, based on the first power characteristic map, the offset is calculated by using a correlation matrix method to obtain the offset of the first power characteristic map, specifically: the channel data at each position of the first power characteristic map is converted into a power characteristic vector to obtain a power characteristic vector at each position; by calculating the cosine similarity between each power characteristic vector in the first power characteristic map, the correlation value between the power characteristic vectors is obtained and a correlation matrix is ​​constructed; the correlation value between each power characteristic vector is compared with a preset threshold value, and the position pairs with higher correlation in the correlation matrix are extracted; the coordinate difference of the position pairs with higher correlation in the correlation matrix is ​​calculated to obtain the offsets in height and width respectively; In a preferred embodiment, the power factor prediction value of each reactive compensation node is input into a preset reactive power adjustment calculation formula to obtain a reactive power error value and a reactive power adjustment value, specifically: based on the power factor prediction value and the actual power factor value of each reactive compensation node, a reactive power error value is obtained; the reactive power error value is input into a preset reactive power adjustment calculation formula to obtain a reactive power adjustment value.

[0008] In a preferred embodiment, the reactive power error value is graded and the reactive power of each reactive compensation node is corrected based on the reactive power adjustment value, specifically: when the reactive power error value is greater than a preset maximum error threshold, it is divided into a first-level error, and a larger adjustment weight is used to prioritize reactive compensation; when the reactive power error value is less than a preset maximum error threshold and greater than a preset minimum error, it is divided into a second-level error, and a medium adjustment weight is used for conventional reactive compensation; when the reactive power error value is less than a preset minimum error threshold, it is divided into a third-level error, and a smaller adjustment weight is used for minor reactive compensation.

[0009] The device of any one of the methods for centralized coordinated control of distribution networks based on photovoltaic power generation and reactive power compensation is characterized in that it comprises a data acquisition module, a power prediction module, a power adjustment module, and a coordinated control module, and there are connections between the modules: A data acquisition module, which acquires historical photovoltaic power generation data and environmental data of each reactive compensation node in the distribution network and performs preprocessing, wherein the historical photovoltaic power generation data of each reactive compensation node in the distribution network includes voltage data, current data, power data, and reactive compensation node status data; A power prediction module inputs historical photovoltaic power generation data and environmental data of each reactive compensation node of the distribution network into a preset bidirectional LSTM model based on offset adjustment to obtain a power factor prediction value of each reactive compensation node; The power adjustment module inputs the power factor prediction value of each reactive power compensation node into a preset reactive power adjustment calculation formula to obtain a reactive power error value and a reactive power adjustment value; The collaborative control module classifies the reactive power error value and corrects the reactive power of each reactive compensation node based on the reactive power adjustment value.

[0010] An electronic device, characterized in that the electronic device comprises: at least one processor; and, a memory communicatively connected to the at least one processor; wherein, The memory stores a computer program executable by the at least one processor, The computer program is executed by the at least one processor, So that the at least one processor can execute the centralized coordinated control method of the distribution network based on photovoltaic power generation and reactive power compensation as described in any one of claims 1 to 7.

[0011] A computer-readable storage medium storing a computer program, characterized in that when the computer program is executed by a processor, the centralized collaborative control method for a distribution network based on photovoltaic power generation and reactive power compensation as described in any one of claims 1 to 7 is implemented.

[0012] The technical effects and advantages of the centralized coordinated control method, device, equipment and medium of the distribution network based on photovoltaic power generation and reactive power compensation of the present invention are as follows: 1. The present invention obtains the historical photovoltaic power generation data and environmental data of each reactive compensation node in the distribution network as the input of the power prediction model, which can more accurately reflect the external influencing factors of photovoltaic power generation and improve the prediction accuracy; the power factor prediction value of each reactive compensation node is input into the preset reactive power adjustment calculation formula, and each node is graded and corrected after obtaining the reactive power error value and adjustment value, which helps to accurately locate the deviation source of reactive power and make targeted adjustments. This method can optimize the reactive power distribution of the power grid, reduce power grid losses, improve overall operating efficiency and stability, and at the same time ensure that the reactive compensation of each node is more reasonable and efficient.

[0013] 2. The present invention predicts the power factor of each reactive compensation node by constructing a bidirectional LSTM model based on offset adjustment. First, the bidirectional LSTM model is good at handling long-term dependencies and complex nonlinear relationships, and can accurately model the complex interactions between the power factor of the reactive compensation node and variables such as environmental factors and photovoltaic power generation, thereby improving the accuracy of the prediction; secondly, since the power factor of the distribution network is often affected by dynamic factors such as weather and load changes, the model can adapt to these dynamic changes more flexibly through offset adjustment. It can help the model maintain a high prediction accuracy when environmental conditions such as light intensity and temperature change dramatically, thereby improving the model's ability to cope with complex and changing environments. BRIEF DESCRIPTION OF THE DRAWINGS

[0014] Figure 1 It is a structural schematic diagram of the centralized coordinated control method of the distribution network based on photovoltaic power generation and reactive power compensation of the present invention.

[0015] Figure 2 This is a functional module diagram of the device of the centralized collaborative control method of the distribution network based on photovoltaic power generation and reactive power compensation of the present invention.

[0016] Figure 3 It is a schematic diagram of the electronic device structure of the centralized coordinated control method of the distribution network based on photovoltaic power generation and reactive power compensation of the present invention. DETAILED DESCRIPTION

[0017] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0018] Embodiment 1, Figure 1 The invention provides a centralized coordinated control method for a distribution network based on photovoltaic power generation and reactive power compensation, comprising the following steps: S1, obtaining historical photovoltaic power generation data and environmental data of each reactive compensation node in the distribution network and preprocessing them, wherein the historical photovoltaic power generation data of each reactive compensation node in the distribution network includes voltage data, current data, power data, and reactive compensation node status data; In this example, the environmental data is specifically: The environmental data include light intensity, temperature, humidity, air pressure, and air quality.

[0019] It should be noted that environmental data directly affects the output power of photovoltaic power generation systems. Accurately obtaining this data helps to more accurately predict photovoltaic power generation, thereby improving the accuracy of power factor prediction. The power generation efficiency of photovoltaic modules is significantly affected by environmental conditions. For example, high temperature may reduce power generation efficiency, and changes in light intensity will directly affect power generation; secondly, comprehensive analysis of environmental data can identify and adjust power generation fluctuations caused by weather changes, help optimize reactive power compensation strategies, maintain the stability of the power grid, and effectively respond to power generation fluctuations caused by environmental changes by timely adjusting reactive power compensation.

[0020] In this example, the historical photovoltaic power generation data and environmental data of each reactive power compensation node in the distribution network are obtained and preprocessed, specifically: The historical photovoltaic power generation data and environmental data of each reactive compensation node in the distribution network are cleaned, the missing values ​​in the data are processed, and the historical photovoltaic power generation data and environmental data of each reactive compensation node in the distribution network are normalized using Min-Max standardization; Among them, the Min-Max normalization formula is:

[0021] Among them, x is the historical photovoltaic power generation data and environmental data of each reactive compensation node, , are the minimum and maximum values ​​of the historical photovoltaic power generation data and environmental data of each reactive compensation node, It is the standardized historical photovoltaic power generation data and environmental data of each reactive compensation node; It should be noted that the use of Min-Max standardization to normalize the historical photovoltaic power generation data and environmental data of each reactive compensation node in the distribution network helps to scale the data to a unified range (usually between 0 and 1), thereby eliminating the impact of different dimensions and data ranges on model training and prediction. This normalization method can improve the stability and convergence speed of the model, avoid training imbalances caused by differences in data ranges, ensure the reasonable distribution of weights of different data features in the model, and improve the accuracy and reliability of predictions. In addition, Min-Max standardization makes the data more suitable for various machine learning algorithms and optimizes the performance of the model.

[0022] S2, inputting the historical photovoltaic power generation data and environmental data of each reactive compensation node of the distribution network into a preset bidirectional LSTM model based on offset adjustment to obtain a power factor prediction value of each reactive compensation node; In this example, the preset bidirectional LSTM model based on offset adjustment is specifically: Constructing a convolution layer to extract features from historical photovoltaic power generation data and environmental data of each reactive compensation node in the distribution network to obtain a first power feature graph; Based on the first power characteristic graph, an offset is calculated using a correlation matrix method to obtain an offset of the first power characteristic graph; Inputting the offset of the first power characteristic map and the first power characteristic map into a preset variability convolution module for feature extraction to obtain a second power characteristic map; The second power characteristic graph is used as the input of the preset bidirectional LSTM model to obtain a bidirectional LSTM model based on offset adjustment.

[0023] In this example, the convolution layer is constructed to extract features from the historical photovoltaic power generation data and environmental data of each reactive compensation node in the distribution network to obtain a first power feature graph, specifically: Set the size of the convolution kernel to 3x3 and the number to 32 By constructing a convolutional layer, the historical photovoltaic power generation data and environmental data of each reactive compensation node in the distribution network are feature extracted. First, the data is normalized and converted into a format suitable for input into the convolutional network, and then the convolution kernel is applied to the data to perform a convolution operation to extract local features. By setting the appropriate convolution kernel size, stride, and padding, multiple feature maps are generated, and nonlinear transformations are introduced using activation functions. Finally, these feature maps are integrated to form the first power feature map, which captures the key patterns and relationships in the data and provides rich feature information for subsequent analysis.

[0024] The power factor prediction value of each reactive compensation node is input into a preset reactive power adjustment calculation formula to obtain a reactive power error value and a reactive power adjustment value; the reactive power error value is graded and the reactive power of each reactive compensation node is corrected based on the reactive power adjustment value.

[0025] In this example, based on the first power characteristic graph, the offset is calculated by using a correlation matrix method to obtain the offset of the first power characteristic graph, specifically: Converting the channel data at each position of the first power characteristic graph into a power characteristic vector to obtain a power characteristic vector at each position; By calculating the Person correlation coefficient between each power feature vector in the first power feature graph, the correlation value between the power feature vectors is obtained and a correlation matrix is ​​constructed; Compare the correlation value between each power feature vector with a preset threshold value, and extract position pairs with higher correlation in the correlation matrix; Calculate the coordinate differences of the position pairs with higher correlation in the correlation matrix to obtain the offsets in height and width respectively; In this example, the channel data at each position of the first power characteristic graph is converted into a power characteristic vector to obtain a power characteristic vector at each position, specifically: Pull the C channel data at each position of the first power characteristic graph into a vector to obtain the characteristic vector of each position. The specific formula is as follows:

[0026] where , represents the eigenvalue of the i-th channel at position (h,w), is the power eigenvector for each position.

[0027] In this example, the cosine similarity between each power feature vector in the first power feature graph is calculated to obtain the correlation value between the power feature vectors and construct a correlation matrix, specifically: For two locations ( )and , the correlation between them can be obtained by calculating the Person correlation coefficient, where the calculation formula is as follows:

[0028] in, represents the covariance, , are the standard deviations at two locations, It is an NxN correlation matrix, each element of which represents the correlation of different spatial positions of the feature map.

[0029] It should be noted that based on the high correlation areas in the correlation matrix, the offset between these areas can be inferred. For example, if two different locations , The correlation is very high, which means that their features are very similar at different spatial locations. Therefore, it can be inferred that some spatial shift has occurred in the data. By analyzing the correlation matrix, we can find out which spatial positions have strong correlations, and thus infer which features have obvious shift trends in time or space.

[0030] In this example, the correlation value between each power feature vector is compared with a preset threshold value to extract position pairs with higher correlation in the correlation matrix, specifically: By setting the threshold, the position pairs with higher correlation in the correlation matrix are extracted, specifically: in is the preset threshold.

[0031] Traverse each element in the correlation matrix R to determine whether its value is greater than a threshold. If it is greater than the threshold, it indicates that there is a high correlation between the two positions. Then, all position pairs that meet the above conditions are stored to form a set of high-correlation position pairs.

[0032] It should be noted that the purpose of setting the threshold is to filter out position pairs with lower correlation and retain those with higher correlation. It is usually set to 0.8, and only position pairs with correlation greater than 0.8 are considered.

[0033] In this example, the coordinate differences of the position pairs with higher correlation in the correlation matrix are calculated to obtain the offsets in height and width, respectively, specifically: For each pair of highly correlated positions, calculate the The specific formula for the offset in the feature map is as follows:

[0034] in, and Represents the offset in height and width directions respectively.

[0035] The above-mentioned offsets in the height and width directions are then stored to form an offset mapping table, where each high-correlation position pair corresponds to an offset.

[0036] In this example, the offset of the first power characteristic map and the first power characteristic map are input into a preset variability convolution module for feature extraction to obtain a second power characteristic map, specifically:

[0037]

[0038] in, For location The offset of is the convolution kernel weight that generates the offset, is the weight of the convolutional layer, is the convolution operation, The second power characteristic diagram is The value of the position, Relative to the center position The offset of is the set of sampling points within the convolution window, The convolution kernel is The weight of the position, is the first power characteristic diagram, is the center position of the feature map.

[0039] S3, inputting the power factor prediction value of each reactive power compensation node into a preset reactive power adjustment calculation formula to obtain a reactive power error value and a reactive power adjustment value; In this example, the power factor prediction value of each reactive power compensation node is input into a preset reactive power adjustment calculation formula to obtain a reactive power error value and a reactive power adjustment value, specifically: Obtaining a reactive power error value based on the power factor prediction value and the actual power factor value of each reactive power compensation node; The reactive power error value is input into a preset reactive power adjustment calculation formula to obtain a reactive power adjustment value.

[0040] In this example, the preset reactive power adjustment calculation formula is specifically:

[0041]

[0042] in, is the reactive power error value of the i-th node, is the reactive power prediction value of the ith node, is the actual value of reactive power of the ith node, is the reactive power adjustment value of the i-th node, is the system proportionality coefficient, is the differential coefficient.

[0043] If the calculated That is, if the reactive power adjustment value of the i-th node is positive, it means that the reactive power needs to be increased (usually by enabling the capacitor bank or increasing the output of the reactive compensation device); if it is negative, it means that the reactive power needs to be reduced (by shutting down the capacitor bank or reducing the output of the reactive compensation device).

[0044] It should be noted that the role of applying reactive power adjustment value is to ensure that reactive power in the power system remains balanced, thereby maintaining voltage stability and ensuring safe operation of the power grid. The reactive power adjustment value is calculated based on the error between the actual measured reactive power and the target value. By acting on reactive compensation equipment such as capacitor banks, reactors, static VAR compensators (SVCs) or static synchronous compensators (STATCOMs), real-time dynamic adjustment of reactive power is achieved. When the reactive power of the power grid is insufficient, the system needs to increase reactive power to maintain the voltage level; conversely, when the reactive power is excessive, the reactive power needs to be reduced to prevent excessive voltage. Reasonable application of reactive power adjustment value can not only effectively eliminate the reactive power deviation of the system, but also avoid the problem of grid instability caused by over-compensation or under-compensation. In addition, through the closed-loop feedback control mechanism, reactive power adjustment can be continuously optimized according to factors such as load changes and photovoltaic power generation, so that the power grid can respond to fluctuations more intelligently and flexibly, and improve the overall operation efficiency and reliability of the power grid.

[0045] S4, classifying the reactive power error value and correcting the reactive power of each reactive compensation node based on the reactive power adjustment value.

[0046] In this example, the reactive power error values ​​are classified and the reactive power of each reactive compensation node is corrected based on the reactive power adjustment value, specifically: When the reactive power error value is greater than a preset maximum error threshold, it is divided into a primary error, and reactive power compensation is performed preferentially with a larger adjustment weight; When the reactive power error value is less than a preset maximum error threshold and greater than a preset minimum error, it is divided into a secondary error and conventional reactive power compensation is performed using a medium adjustment weight; When the reactive power error value is less than a preset minimum error threshold, it is divided into three levels of error, and a smaller adjustment weight is used to perform a slight reactive power compensation.

[0047] It should be noted that the first-level error represents a large reactive power gap, which may cause drastic fluctuations in the grid voltage and even lead to grid instability. Therefore, the system should quickly and significantly compensate for reactive power and use a larger adjustment weight to ensure that the voltage returns to a stable state as soon as possible; Second level error means the error is within the acceptable range and the system is not in urgent danger, but conventional reactive power compensation is still required. A medium adjustment weight is used to enable the power grid to maintain stable operation under normal conditions and prevent the error from continuing to increase; The third level error means that the reactive power of the system is basically balanced and the voltage is relatively stable. Only slight reactive power compensation is needed to avoid the shock caused by excessive adjustment. Therefore, a small adjustment weight is used for small compensation to ensure that the system continues to remain stable.

[0048] Among them, for the first-level error, reactive power adjustment can be carried out by quickly putting into use a large-capacity capacitor group or reactor to make the system quickly restore stability; the second-level error is slowly adjusted through conventional reactive compensation equipment, for example, gradually putting into use or withdrawing some reactive compensation equipment; the third-level error can be compensated slightly by adjusting the reactive output of the static VAR compensator (SVC) or the static synchronous compensator (STATCOM).

[0049] In addition, the adjustment weight determines the intensity of compensation according to the error level. Level 1 error means that the system is in an unstable state and requires significant intervention, so the weight is larger; level 2 error indicates that the system needs moderate adjustment; level 3 error means that the system is close to a stable state and only requires slight adjustment. The difference in weights ensures that the system can respond quickly to major errors while avoiding over-adjustment under minor errors, ensuring the accuracy of compensation.

[0050] like Figure 2 As shown, it is a functional module diagram of a centralized collaborative control method device for distribution networks based on photovoltaic power generation and reactive power compensation. According to the functions to be implemented, the centralized collaborative control device for distribution networks based on photovoltaic power generation and reactive power compensation may include a data acquisition module, a power prediction module, a power adjustment module, and a collaborative control module. The module described in the present invention may also be referred to as a unit, which refers to a series of computer program segments that can be executed by a processor of a personnel performance data processing device and can complete fixed functions, which are stored in a memory of the personnel performance data processing device.

[0051] In this embodiment, the functions of each module are as follows: A data acquisition module, which acquires historical photovoltaic power generation data and environmental data of each reactive compensation node in the distribution network and performs preprocessing, wherein the historical photovoltaic power generation data of each reactive compensation node in the distribution network includes voltage data, current data, power data, and reactive compensation node status data; A power prediction module inputs historical photovoltaic power generation data and environmental data of each reactive compensation node of the distribution network into a preset bidirectional LSTM model based on offset adjustment to obtain a power factor prediction value of each reactive compensation node; The power adjustment module inputs the power factor prediction value of each reactive power compensation node into a preset reactive power adjustment calculation formula to obtain a reactive power error value and a reactive power adjustment value; The collaborative control module classifies the reactive power error value and corrects the reactive power of each reactive compensation node based on the reactive power adjustment value.

[0052] In detail, each module described in the centralized coordinated control device for distribution network based on photovoltaic power generation and reactive power compensation described in the embodiment of the present invention adopts the same technical means as the centralized coordinated control method for distribution network based on photovoltaic power generation and reactive power compensation described in the attached drawings when used, and can produce the same technical effect, which will not be repeated here. like Figure 3 As shown, it is a structural schematic diagram of an electronic device of the centralized coordinated control method of the distribution network based on photovoltaic power generation and reactive power compensation of the present invention.

[0053] The electronic device may include a processor, a memory, a communication bus and a communication interface, and may also include a computer program stored in the memory and executable on the processor, such as a centralized collaborative control program for a distribution network based on photovoltaic power generation and reactive power compensation.

[0054] Among them, the processor is the control core (Control Unit) of the electronic device, which uses various interfaces and lines to connect the various components of the entire electronic device, and executes various functions of the electronic device and processes data by running or executing programs or modules stored in the memory and calling data stored in the memory.

[0055] The memory includes at least one type of readable storage medium, and in some embodiments, the memory may be an internal storage unit of the electronic device, such as a mobile hard disk of the electronic device. The memory may be used to store not only application software installed in the electronic device but also various data.

[0056] The communication bus is configured to realize connection and communication between the memory and at least one processor.

[0057] The communication interface is used for communication between the above electronic device and other devices, including a network interface and a user interface.

[0058] The figure only shows an electronic device with components. Those skilled in the art will understand that the structure shown in the figure does not constitute a limitation on the electronic device, and may include fewer or more components than shown in the figure, or combine certain components, or arrange the components differently.

[0059] It should be understood that the embodiment is for illustration only and the scope of the patent application is not limited to this structure.

[0060] The centralized collaborative control program for distribution networks based on photovoltaic power generation and reactive power compensation stored in the memory in the electronic device is a combination of multiple instructions. When running in the processor, the steps in the above-mentioned method for centralized collaborative control of distribution networks based on photovoltaic power generation and reactive power compensation can be implemented.

[0061] Specifically, the specific implementation system of the processor for the above instructions can refer to the description of the relevant steps in the corresponding embodiment of the accompanying drawings, which will not be repeated here.

[0062] The present invention also provides a computer-readable storage medium, which stores a computer program. When the computer program is executed by a processor of an electronic device, it can implement the steps in the above-mentioned centralized collaborative control method of the distribution network based on photovoltaic power generation and reactive power compensation.

[0063] The above formulas are all dimensionless and numerical calculations. The formula is a formula for the most recent real situation obtained by collecting a large amount of data and performing software simulation. The preset parameters in the formula are set by technicians in this field according to actual conditions.

[0064] The above embodiments may be implemented in whole or in part by software, hardware, firmware or any other combination. When implemented by software, the above embodiments may be implemented in whole or in part in the form of a computer program product.

[0065] Those of ordinary skill in the art will appreciate that the modules and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of this application.

[0066] In addition, each functional module in each embodiment of the present application may be integrated into one processing module, or each module may exist physically separately, or two or more modules may be integrated into one module.

[0067] The above is only a specific implementation of the present application, but the protection scope of the present application is not limited thereto. Any person skilled in the art who is familiar with the present technical field can easily think of changes or substitutions within the technical scope disclosed in the present application, which should be included in the protection scope of the present application. Therefore, the protection scope of the present application should be based on the protection scope of the claims.

[0068] Finally: The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the protection scope of the present invention.

Claims

1. A centralized coordinated control method for distribution network based on photovoltaic power generation and reactive power compensation, characterized in that: The following steps are involved: Acquire historical photovoltaic power generation data and environmental data of each reactive compensation node of the distribution network and perform preprocessing, wherein the historical photovoltaic power generation data of each reactive compensation node of the distribution network includes voltage data, current data, power data, and reactive compensation node status data; Inputting historical photovoltaic power generation data and environmental data of each reactive compensation node of the distribution network into a preset bidirectional LSTM model based on offset adjustment to obtain a power factor prediction value of each reactive compensation node; Inputting the power factor prediction value of each reactive power compensation node into a preset reactive power adjustment calculation formula to obtain a reactive power error value and a reactive power adjustment value; The reactive power error value is classified and the reactive power of each reactive compensation node is corrected based on the reactive power adjustment value.

2. The centralized coordinated control method of distribution network based on photovoltaic power generation and reactive power compensation according to claim 1 is characterized in that: The preset bidirectional LSTM model based on offset adjustment is specifically: Constructing a convolutional layer to extract features from historical photovoltaic power generation data and environmental data of each reactive compensation node in the distribution network to obtain a first power feature graph; Based on the first power characteristic graph, an offset is calculated using a correlation matrix method to obtain an offset of the first power characteristic graph; Inputting the offset of the first power characteristic map and the first power characteristic map into a preset variability convolution module for feature extraction to obtain a second power characteristic map; The second power characteristic graph is used as the input of the preset bidirectional LSTM model to obtain a bidirectional LSTM model based on offset adjustment.

3. The centralized coordinated control method of distribution network based on photovoltaic power generation and reactive power compensation according to claim 2 is characterized in that: Based on the first power characteristic graph, the offset is calculated by using the correlation matrix method to obtain the offset of the first power characteristic graph, which is specifically: Converting the channel data at each position of the first power characteristic graph into a power characteristic vector to obtain a power characteristic vector at each position; By calculating the cosine similarity between each power feature vector in the first power feature graph, the correlation value between the power feature vectors is obtained and a correlation matrix is ​​constructed; Compare the correlation value between each power feature vector with a preset threshold value, and extract position pairs with higher correlation in the correlation matrix; The coordinate differences of the position pairs with higher correlation in the correlation matrix are calculated to obtain the offsets in height and width respectively.

4. The centralized coordinated control method of distribution network based on photovoltaic power generation and reactive power compensation according to claim 1 is characterized in that: The power factor prediction value of each reactive compensation node is input into the preset reactive power adjustment calculation formula to obtain the reactive power error value and the reactive power adjustment value, which are specifically: Obtaining a reactive power error value based on the power factor prediction value and the actual power factor value of each reactive power compensation node; The reactive power error value is input into a preset reactive power adjustment calculation formula to obtain a reactive power adjustment value.

5. The centralized coordinated control method of distribution network based on photovoltaic power generation and reactive power compensation according to claim 1 is characterized in that: The reactive power error value is graded and the reactive power of each reactive compensation node is corrected based on the reactive power adjustment value, specifically: When the reactive power error value is greater than a preset maximum error threshold, it is divided into a primary error, and reactive power compensation is performed preferentially with a larger adjustment weight; When the reactive power error value is less than a preset maximum error threshold and greater than a preset minimum error, it is divided into a secondary error and conventional reactive power compensation is performed using a medium adjustment weight; When the reactive power error value is less than a preset minimum error threshold, it is divided into three levels of error, and a smaller adjustment weight is used to perform a slight reactive power compensation.

6. The centralized coordinated control method of distribution network based on photovoltaic power generation and reactive power compensation according to claim 2 is characterized in that: The offset of the first power feature map and the first power feature map are input into a preset variability convolution module for feature extraction to obtain a second power feature map, specifically: in, For location The offset of is the convolution kernel weight that generates the offset, is the weight of the convolutional layer, is the convolution operation, The second power characteristic diagram is The value of the position, Relative to the center position The offset of is the set of sampling points within the convolution window, The convolution kernel is The weight of the position, is the first power characteristic diagram, is the center position of the feature map.

7. The centralized coordinated control method for distribution network based on photovoltaic power generation and reactive power compensation according to claim 3 is characterized in that: The preset reactive power adjustment calculation formula is specifically: in, is the reactive power error value of the i-th node, is the reactive power prediction value of the ith node, is the actual value of reactive power of the ith node, is the reactive power adjustment value of the i-th node, is the system proportionality coefficient, is the differential coefficient.

8. A device using the centralized coordinated control method for distribution network based on photovoltaic power generation and reactive power compensation as described in any one of claims 1 to 7, characterized in that: It includes data acquisition module, power prediction module, power adjustment module and collaborative control module. There are connections between the modules: A data acquisition module, which acquires historical photovoltaic power generation data and environmental data of each reactive compensation node in the distribution network and performs preprocessing, wherein the historical photovoltaic power generation data of each reactive compensation node in the distribution network includes voltage data, current data, power data, and reactive compensation node status data; A power prediction module inputs historical photovoltaic power generation data and environmental data of each reactive compensation node of the distribution network into a preset bidirectional LSTM model based on offset adjustment to obtain a power factor prediction value of each reactive compensation node; The power adjustment module inputs the power factor prediction value of each reactive power compensation node into a preset reactive power adjustment calculation formula to obtain a reactive power error value and a reactive power adjustment value; The collaborative control module classifies the reactive power error value and corrects the reactive power of each reactive compensation node based on the reactive power adjustment value.

9. An electronic device, characterized in that: The electronic device comprises: at least one processor; and, a memory communicatively connected to the at least one processor; wherein, The memory stores a computer program executable by the at least one processor, The computer program is executed by the at least one processor, So that the at least one processor can execute the centralized coordinated control method of the distribution network based on photovoltaic power generation and reactive power compensation as described in any one of claims 1 to 7.

10. A computer-readable storage medium storing a computer program, characterized in that: When the computer program is executed by a processor, the centralized coordinated control method of the distribution network based on photovoltaic power generation and reactive power compensation as described in any one of claims 1 to 7 is implemented.

Images