Day-ahead discrete reactive power device action planning method and device based on prediction information

Abstract

The application discloses a kind of day-ahead discrete reactive power equipment action planning method and device based on prediction information.According to the main influencing factor of target node voltage determined according to power grid historical operation data;According to the main influencing factor, the corresponding data of each factor is selected from the power grid historical operation data;The data selected is used to train Attention-LSTM model, and the voltage operation trend prediction model is obtained;The day-ahead prediction information data of each influencing factor is input into the voltage operation trend prediction model, and the day-ahead voltage operation trend is predicted;According to day-ahead voltage operation trend, day-ahead discrete reactive power equipment action planning is carried out.The application can flexibly divide day-ahead voltage control period and the number of actions of each segment according to power grid prediction data, effectively avoid the problem that traditional segmentation method cannot respond to power grid change flexibly, strategy lag and need to rely on relevant personnel's work experience, so as to improve the operation level of power grid voltage management.

Description

Technical Field

[0001] This invention relates to the field of power system operation control technology, specifically to a method and apparatus for day-ahead discrete reactive power equipment action planning based on predictive information. Background Technology

[0002] With the increasing penetration of new energy sources and the large-scale construction of DC transmission lines, the power grid faces significant uncertainties, leading to greater voltage fluctuations and increased randomness. Traditional automatic voltage control systems are posing increasing challenges. Currently, capacitive reactors remain the primary means of voltage regulation in power systems. However, for safety and economic reasons, the number of daily operations for capacitive reactors is limited, requiring manual allocation of operation times and frequencies based on experience. This method not only relies heavily on the experience of personnel but also struggles to respond promptly to changes in new energy sources and DC transmission. With the development of new energy forecasting and load forecasting technologies, and the improvement of various power grid forecasting data, proactive voltage control strategies based on forecast data have become possible. Therefore, it is necessary to develop a method that can flexibly configure the control time and number of operations for day-ahead discrete reactive power equipment (DVP) based on power grid forecasting information to improve the efficiency and accuracy of DVP operation planning. This will ultimately improve voltage control efficiency, reduce voltage fluctuations, and enhance power quality. Summary of the Invention

[0003] The purpose of this invention is to provide a method and apparatus for day-ahead discrete reactive power equipment operation planning based on predictive information, so as to solve the problem that the existing discrete reactive power equipment operation planning needs to rely on manual experience to set, and cannot respond to changes in new energy sources and DC power in a timely manner.

[0004] To achieve the above objectives, the present invention adopts the following technical solution:

[0005] On the one hand, a day-ahead discrete reactive power equipment action planning method based on predictive information includes:

[0006] Based on the pre-determined main influencing factors of the target node voltage, obtain the day-ahead forecast information data corresponding to each factor;

[0007] Input the day-ahead forecast information data corresponding to each factor into the voltage operation trend prediction model to obtain the day-ahead voltage operation trend prediction data;

[0008] Based on the day-ahead voltage operating trend prediction data, the day-ahead discrete reactive power equipment operation planning is performed.

[0009] Furthermore, the main influencing factors of the target node voltage are determined according to the following method:

[0010] Collect historical power grid operation data for the target area, including historical power grid operation data of the target node, its subordinate nodes, and adjacent nodes;

[0011] Calculate the Pearson correlation coefficients between various data and the target node voltage;

[0012] Data types with correlation coefficients exceeding a set value are selected as the main influencing factors of the target node voltage.

[0013] Furthermore, the voltage operating trend prediction model is obtained through the following method:

[0014] By combining the Long Short-Term Memory (LSTM) network model with the attention mechanism model, we obtain the Attention-LSTM model.

[0015] Based on the main influencing factors of the target node voltage, relevant data for each factor are selected from the historical operation data of the power grid in the target area;

[0016] The selected data for each factor are used as input to the Attention-LSTM model, and the target node voltage is used as output to train the Attention-LSTM model, thus obtaining a voltage trend prediction model.

[0017] Further, the step of planning the operation of discrete reactive power equipment based on the day-ahead voltage operating trend prediction data includes:

[0018] The daytime voltage operation trend forecast data is divided into a target number of segments using the trend segmentation method, and the duration of each segment is not less than the preset duration.

[0019] The number of times discrete reactive power equipment operates is allocated based on the voltage fluctuation of each segment.

[0020] Furthermore, the day-ahead voltage operating trend forecast data includes several voltage forecast results, and the method of dividing the day-ahead voltage operating trend forecast data into a target number of segments using the trend segmentation method includes:

[0021] If each voltage prediction result is taken as a segmentation point, the day-ahead voltage operating trend prediction data is divided into several segments.

[0022] Calculate the voltage change between adjacent segments, find the two adjacent segments with the closest voltage change, and merge these two segments; repeat this step until the number of segments reaches the target number and the duration of each segment is not less than the preset duration.

[0023] Furthermore, the process of finding the two adjacent segments with the closest voltage changes is calculated using the following formula:

[0024] ;

[0025] In the formula: Indicates the first Voltage prediction data for each segment point, This represents the voltage trend change between two adjacent segments with the closest change trends among N segments, where N represents the number of segments before each merging.

[0026] Furthermore, while finding and merging two adjacent segments with the closest voltage changes, the voltage fluctuation is calculated according to the following formula:

[0027] ;

[0028] In the formula: Indicates the first Voltage fluctuation in each segment This represents the total voltage fluctuation.

[0029] Furthermore, the allocation of the number of discrete reactive power device operations based on the voltage fluctuation of each segment includes:

[0030] The number of actions of discrete reactive power equipment is allocated by integer, with each segment allocated at least 1 action, and the rest allocated according to the proportion of voltage fluctuation in each segment to the total voltage fluctuation.

[0031] If there are still action counts remaining after integer allocation, then the remaining action counts will be allocated using the remainder.

[0032] Furthermore, the integer allocation formula is as follows:

[0033] ;

[0034] In the formula: Indicates the first Number of actions for each segmented discrete quantity This indicates the total number of actions allowed per day. The integer represents the number of segments. Indicates the first Voltage fluctuation in each segment This represents the total voltage fluctuation.

[0035] Furthermore, the remaining number of actions is allocated using the remainder, which includes:

[0036] Calculate the remainder of each segment after integer distribution using the following formula. :

[0037] ;

[0038] in, Indicates the first Voltage fluctuation in each segment This indicates the total number of actions allowed per day. For the number of segments, This represents the total voltage fluctuation; the percentage is calculated using the remainder.

[0039] Sort the segments in descending order based on the remainders, and increment the number of actions for each segment by 1 according to the sorting order, until the number of actions is allocated.

[0040] On the other hand, a day-ahead discrete reactive power equipment action planning device based on predictive information includes:

[0041] The acquisition module acquires the day-ahead forecast information data corresponding to each of the pre-determined main influencing factors of the target node voltage.

[0042] The prediction module inputs the day-ahead forecast information data corresponding to each factor into the voltage operation trend prediction model to obtain the day-ahead voltage operation trend prediction data.

[0043] The action planning module performs day-ahead discrete reactive power equipment action planning based on the day-ahead voltage operating trend prediction data.

[0044] Furthermore, the aforementioned day-ahead discrete reactive power equipment action planning device based on predictive information further includes: an influencing factor determination module, used for:

[0045] Collect historical power grid operation data for the target area, including historical power grid operation data of the target node, its subordinate nodes, and adjacent nodes;

[0046] Calculate the Pearson correlation coefficients between various data and the target node voltage;

[0047] Data types with correlation coefficients exceeding a set value are selected as the main influencing factors of the target node voltage.

[0048] Furthermore, the voltage operation trend prediction model adopts an Attention-LSTM model, and the device also includes a training module, which is used for:

[0049] Based on the main influencing factors of the target node voltage, relevant data for each factor are selected from the historical operation data of the power grid in the target area;

[0050] The selected data for each factor are used as input to the Attention-LSTM model, and the target node voltage is used as output to train the Attention-LSTM model, thus obtaining a voltage trend prediction model.

[0051] Furthermore, the action planning module includes:

[0052] The segmentation module uses a trend segmentation method to divide the day-ahead voltage operation trend prediction data into a target number of segments, and the duration of each segment is not less than the preset duration.

[0053] The allocation module allocates the number of times discrete reactive power equipment operates based on the voltage fluctuation of each segment.

[0054] Furthermore, the day-ahead voltage operating trend prediction data includes several voltage prediction results;

[0055] The segmentation module is used to divide the day-ahead voltage trend prediction data into several segments by taking each voltage prediction result as a segmentation point; calculate the voltage change between each adjacent segment, find the two adjacent segments with the closest voltage change, merge the two segments, and repeat this step until the number of segments reaches the target number of segments and the duration of each segment is not less than the preset duration.

[0056] Furthermore, the segmentation module finds the two adjacent segments with the closest voltage changes according to the following formula:

[0057] ;

[0058] In the formula: Indicates the first Voltage prediction data for each segment point, This represents the voltage trend change between two adjacent segments with the closest change trends among N segments, where N represents the number of segments before each merging.

[0059] Furthermore, the action planning module also includes a voltage fluctuation statistics module, used for:

[0060] While the segmentation module finds and merges two adjacent segments with the closest voltage changes, it calculates the voltage fluctuation using the following formula:

[0061] ;

[0062] In the formula: Indicates the first Voltage fluctuation in each segment This represents the total voltage fluctuation.

[0063] Furthermore, the allocation module includes:

[0064] The integer allocation module allocates the number of actions of discrete reactive power equipment to integers, with each segment allocated at least one action, and the rest allocated according to the proportion of voltage fluctuation in each segment to the total voltage fluctuation.

[0065] The remainder allocation module allocates the remaining action counts if there are still action counts remaining after integer allocation.

[0066] Furthermore, the integer allocation module allocates according to the following formula:

[0067] ;

[0068] In the formula: Indicates the first Number of actions for each segmented discrete quantity This indicates the total number of actions allowed per day. The integer represents the number of segments. Indicates the first Voltage fluctuation in each segment This represents the total voltage fluctuation.

[0069] Furthermore, the remainder allocation module is used for:

[0070] Calculate the remainder of each segment after integer distribution using the following formula. :

[0071] ;

[0072] Indicates the first Voltage fluctuation in each segment This indicates the total number of actions allowed per day. For the number of segments, This represents the total voltage fluctuation; the percentage is calculated using the remainder.

[0073] Sort the segments in descending order based on the remainders, and increment the number of actions for each segment by 1 according to the sorting order, until the number of actions is allocated.

[0074] An electronic device includes one or more processors, a memory, and one or more programs, wherein the one or more programs are stored in the memory and configured to be executed by the one or more processors, and the one or more programs include instructions for performing any of the methods.

[0075] A readable storage medium, wherein the one or more programs include instructions that, when executed by a computing device, cause the computing device to perform any of the methods described in the method.

[0076] Compared with the prior art, the beneficial technical effects of the present invention are as follows:

[0077] This invention can flexibly divide day-ahead voltage control periods and the number of actions per period based on grid forecast data, effectively avoiding the problems of traditional segmentation methods that cannot flexibly respond to grid changes, have lagging strategies, and rely on the professional experience of relevant personnel, thereby improving the level of grid voltage management and operation. By introducing neural networks and new energy and load forecasting methods, voltage operation trends are predicted, and day-ahead reactive power equipment operation planning is performed based on the prediction results. This helps the automatic voltage control system respond promptly to changes in new energy, DC, and load data, improving the accuracy of grid voltage control, enhancing power quality, and improving grid security. Attached Figure Description

[0078] Figure 1 This is a flowchart of the method of the present invention;

[0079] Figure 2 This is a flowchart of the current discrete reactive power equipment operation planning process. Detailed Implementation

[0080] The present invention will be further described below with reference to specific embodiments. These embodiments are only used to more clearly illustrate the technical solutions of the present invention and should not be construed as limiting the scope of protection of the present invention.

[0081] like Figure 1 As shown, a day-ahead discrete reactive power equipment action planning method based on predictive information includes:

[0082] Step S1: Based on the pre-determined main influencing factors of the target node voltage, obtain the day-ahead forecast information data corresponding to each factor;

[0083] The main influencing factors of the target node voltage were determined using the following methods:

[0084] Collect historical power grid operation data for the target area, including historical power grid operation data of the target node, its subordinate nodes, and adjacent nodes;

[0085] Calculate the Pearson correlation coefficients between various data and the target node voltage;

[0086] Data types with correlation coefficients exceeding a set value are selected as the main influencing factors of the target node voltage.

[0087] The target node refers to the substation node to be optimized.

[0088] Step S2: Input the day-ahead forecast information of each factor into the voltage operation trend forecast model to obtain the day-ahead voltage operation trend forecast data;

[0089] Among them, the voltage operation trend prediction model is built based on the Attention-LSTM neural network model.

[0090] Specifically, the voltage operation trend prediction model is obtained through the following methods:

[0091] By combining the LSTM model with the Attention model, we obtain the Attention-LSTM model;

[0092] Based on the main influencing factors of the target node voltage, relevant data for each factor are selected from the historical operation data of the power grid in the target area;

[0093] The selected data for each factor are used as input to the Attention-LSTM model, and the target node voltage is used as output to train the Attention-LSTM model, thus obtaining a voltage trend prediction model.

[0094] Step S3: Perform day-ahead discrete reactive power equipment operation planning based on the day-ahead voltage operation trend prediction data.

[0095] Step S31: The daytime voltage operation trend prediction data is divided into a target number of segments using the trend segmentation method, and the duration of each segment is not less than the preset duration.

[0096] The day-ahead voltage trend forecast data includes several voltage forecast results. The trend segmentation method is used to divide the day-ahead voltage trend forecast data into a target number of segments, including:

[0097] If each voltage prediction result is taken as a segmentation point, the day-ahead voltage operating trend prediction data is divided into several segments.

[0098] Calculate the voltage change between adjacent segments, find the two adjacent segments with the closest voltage change, and merge these two segments; repeat this step until the number of segments reaches the target number and the duration of each segment is not less than the preset duration.

[0099] Among them, the two adjacent segments with the closest voltage changes are found and calculated according to the following formula:

[0100] ;

[0101] In the formula: Indicates the first Voltage prediction data for each segment point, This represents the voltage trend change between two adjacent segments with the closest change trends among N segments, where N represents the number of segments before each merging.

[0102] While merging two adjacent segments with the closest voltage changes, the voltage fluctuation is calculated using the following formula:

[0103] ;

[0104] In the formula: Indicates the first Voltage fluctuation in each segment This represents the total voltage fluctuation.

[0105] Step S32: Allocate the number of operations for discrete reactive power equipment based on the voltage fluctuation of each segment.

[0106] The number of actions of discrete equipment is allocated based on the voltage fluctuation of each segment. Since the number of actions can only be rounded up and each segment has at least one action, in order to ensure that the number of actions is fully allocated, it is necessary to carry out the allocation in two steps: integer allocation and margin allocation.

[0107] Step S321: Allocate the number of actions of the discrete reactive power equipment to integers, with each segment allocated at least 1 action, and the rest allocated according to the proportion of the voltage fluctuation of each segment to the total voltage fluctuation.

[0108] The integer allocation formula is as follows:

[0109] ;

[0110] In the formula: Indicates the first Number of actions for each segmented discrete quantity This indicates the total number of actions allowed per day. The integer represents the number of segments. Indicates the first Voltage fluctuation in each segment This represents the total voltage fluctuation.

[0111] Step S322: If there are still action counts remaining after the integer allocation is completed, then the remaining action counts are allocated using the remainder.

[0112] Calculate the remainder of each segment after integer distribution using the following formula. :

[0113] ;

[0114] in, Indicates the first Voltage fluctuation in each segment This indicates the total number of actions allowed per day. For the number of segments, This represents the total voltage fluctuation; the percentage is calculated using the remainder.

[0115] Sort the segments in descending order based on the remainders, and increment the number of actions for each segment by 1 according to the sorting order, until the number of actions is allocated.

[0116] Example

[0117] Collect historical operating data of the power grid for the past year (sampling period is 15 minutes), including: voltage, active and reactive data of the target area (including target node, subordinate nodes of the target node, and adjacent nodes of the target node), active and reactive power generation data of new energy sources in the area, active and reactive power generation and power generation plans of traditional generator units in the area, active and reactive loads under the area, DC active and reactive data in the area, parameters such as power grid operation mode, and DC landing point near the area also includes DC input active power, etc.

[0118] Based on the power grid operation mode, the historical power grid operation data is grouped, and a segment is extracted from each group to calculate the Pearson correlation coefficient between each data type and the target node voltage. The calculation method is as follows:

[0119] ;

[0120] Where X represents the historical voltage operation data of the target node, and Y is any type of historical power grid operation data other than voltage; Let X and Y be the covariances. , Let X and Y be the standard deviations, respectively. This represents the Pearson correlation coefficient between X and Y.

[0121] Data Y with a correlation coefficient greater than 0.5 were selected as the main influencing factors of the target node voltage.

[0122] The selected data were used to construct a training set and a test set for the model in a 5:1 ratio. The Attention-LSTM model was trained using the training set and tested using the test set to obtain the voltage operation trend prediction model.

[0123] According to a specific implementation method, the main factors affecting the voltage of the target node are identified, including: active and reactive power output of new energy power plants in the region, active and reactive power output of traditional generator units, active and reactive loads under the region, active power input from external DC, power grid operation mode and other major influencing factors; based on these factors, corresponding data are selected from the historical operation data of the power grid, and these selected data are used as model data input, with the target node voltage as the model output, to train the voltage operation trend prediction model.

[0124] After the model is trained, the predicted data for the next 24 hours corresponding to the main influencing factors of the target node voltage are obtained. The predicted data for the next 24 hours corresponding to each influencing factor are input into the voltage operation trend prediction model, and the model outputs the day-ahead voltage operation trend prediction data.

[0125] In one implementation, the predicted data for the next 24 hours corresponding to the main influencing factors of the target node voltage are input into the voltage operation trend prediction model in groups of 15 minutes, and the predicted voltage operation trend data for the next 24 hours is output.

[0126] According to a specific implementation method, the active and reactive power generation data of new energy sources are replaced by the 24-hour predicted output of new energy sources. The active and reactive power generation data of traditional generator sets, the active and reactive load data of the region and the active power data of external DC input are replaced by the 24-hour predicted output of traditional generator sets, the active and reactive load data of the region and the active power data of external DC input. Other data are the real-time operating data of the power grid at the current moment. These data are input into the voltage operation trend prediction model in groups of 15 minutes, and the voltage operation trend prediction data for the previous 24 hours (96 points) is output.

[0127] Then, based on the day-ahead voltage operating trend forecast data, the day-ahead discrete reactive power equipment operation planning is performed, such as... Figure 2 As shown, it specifically includes:

[0128] Step a1: Use the trend segmentation method to divide the day-ahead voltage operation trend forecast data into 4-6 segments, with each segment lasting no less than 2 hours.

[0129] (1) Each voltage prediction result is taken as a segment point. The data sampling period is 15 minutes. Then there are 96 segments in 24 hours, including the first and last segment points, for a total of 97 segment points (0-96).

[0130] (2) Calculate the voltage change between each adjacent segment, denoted as Then, find the two adjacent segments with the closest changes and merge these two segments. The minimum value is given by the following formula:

[0131] ;

[0132] In the formula: Indicates the first Voltage prediction data for each segment point, This represents the voltage trend change between two adjacent segments with the closest trends among the 96 segments. These two segments are then merged, i.e., the segment with the closest trend is deleted. Each segmentation point.

[0133] (3) While performing step (2), the voltage fluctuation is statistically analyzed and used to allocate the number of operations of discrete reactive power equipment. The formula is as follows:

[0134] ;

[0135] In the formula: Indicates the first Voltage fluctuation in each segment This represents the total voltage fluctuation.

[0136] (4) Repeat steps (2) and (3) as the merging process proceeds. The range of values ​​gradually narrows until the number of segments reaches 4-6 and the minimum segment time is not less than 2 hours. Segmentation is then completed and merging ends.

[0137] Step a2: Allocate the number of operations for discrete reactive power equipment based on the voltage fluctuation of each segment.

[0138] The number of actions of discrete reactive power devices is assigned to integers according to the following formula:

[0139]

[0140] In the formula: Indicates the first Number of actions of each segmented discrete device. Indicates the first Voltage fluctuation in each segment This indicates the total number of actions allowed per day. The number of segments is represented by `int`, which indicates integer division.

[0141] If there are still action counts remaining after the integer allocation is completed, that is... If so, then surplus allocation is required.

[0142] First, calculate the remainder of each segment after integer distribution. The formula is as follows:

[0143] ;

[0144] Where % represents the remainder calculation;

[0145] Sort the segments in descending order based on the remainders, and increment the number of actions for each segment by 1 in that order. Until the number of actions is allocated, that is... .

[0146] This allows for the acquisition of the current-day operational plan for discrete reactive power equipment, including the timing and frequency of operations. This plan can serve as the basis for the control strategy of discrete reactive power equipment in an automatic voltage control system, thereby helping the automatic voltage control system to propose a more accurate and safer voltage control strategy that can dynamically respond to changes in new energy sources, DC power, and load.

[0147] The present invention provides a day-ahead discrete reactive power equipment operation planning method based on predictive information. By introducing neural networks and new energy and load prediction methods, voltage operation trend is predicted, and day-ahead reactive power equipment operation planning is performed based on the predicted information. This helps the automatic voltage control system respond promptly to changes in new energy, DC, and load data, improves the accuracy of grid voltage control, and enhances power quality and grid safety.

[0148] In another embodiment, a day-ahead discrete reactive power equipment action planning device based on predictive information includes:

[0149] The acquisition module acquires the day-ahead forecast information data corresponding to each of the pre-determined main influencing factors of the target node voltage.

[0150] The prediction module inputs the day-ahead forecast information data corresponding to each factor into the voltage operation trend prediction model to obtain the day-ahead voltage operation trend prediction data.

[0151] The action planning module performs day-ahead discrete reactive power equipment action planning based on the day-ahead voltage operating trend prediction data.

[0152] Furthermore, the device also includes an influencing factor determination module, used for:

[0153] Collect historical power grid operation data for the target area, including historical power grid operation data of the target node, its subordinate nodes, and adjacent nodes;

[0154] Calculate the Pearson correlation coefficients between various data and the target node voltage;

[0155] Data types with correlation coefficients exceeding a set value are selected as the main influencing factors of the target node voltage.

[0156] The voltage operation trend prediction model adopts the Attention-LSTM model, and the device also includes a training module, which is used for:

[0157] Based on the main influencing factors of the target node voltage, relevant data for each factor are selected from the historical operation data of the power grid in the target area;

[0158] The selected data for each factor are used as input to the Attention-LSTM model, and the target node voltage is used as output to train the Attention-LSTM model, thus obtaining a voltage trend prediction model.

[0159] The motion planning module includes:

[0160] The segmentation module uses a trend segmentation method to divide the day-ahead voltage operation trend prediction data into a target number of segments, and the duration of each segment is not less than the preset duration.

[0161] The allocation module allocates the number of times discrete reactive power equipment operates based on the voltage fluctuation of each segment.

[0162] More specifically, the segmented module is used for:

[0163] If each voltage prediction result output by the voltage operation trend prediction model is taken as a segmentation point, then the day-ahead voltage operation trend prediction data is divided into several segments.

[0164] Calculate the voltage change between adjacent segments, find the two adjacent segments with the closest voltage change, and merge these two segments; repeat this step until the number of segments reaches the target number and the duration of each segment is not less than the preset duration.

[0165] The segmentation module uses the following formula to find two adjacent segments with the closest voltage changes:

[0166] ;

[0167] In the formula: Indicates the first Voltage prediction data for each segment point, This represents the voltage trend change between two adjacent segments with the closest change trends among N segments, where N represents the number of segments before each merging.

[0168] Furthermore, the action planning module also includes a voltage fluctuation statistics module, used for:

[0169] While the segmentation module finds and merges two adjacent segments with the closest voltage changes, it calculates the voltage fluctuation using the following formula:

[0170] ;

[0171] In the formula: Indicates the first Voltage fluctuation in each segment This represents the total voltage fluctuation.

[0172] The allocation module includes:

[0173] The integer allocation module allocates the number of actions of discrete reactive power equipment to integers, with each segment allocated at least one action, and the rest allocated according to the proportion of voltage fluctuation in each segment to the total voltage fluctuation.

[0174] The remainder allocation module allocates the remaining action counts if there are still action counts remaining after integer allocation.

[0175] More specifically, the integer allocation module allocates according to the following formula:

[0176]

[0177] In the formula: Indicates the first Number of actions for each segmented discrete quantity This indicates the total number of actions allowed per day. The integer represents the number of segments. Indicates the first Voltage fluctuation in each segment This represents the total voltage fluctuation.

[0178] More specifically, the remainder allocation module is used for:

[0179] Calculate the remainder of each segment after integer distribution using the following formula. :

[0180] ;

[0181] Sort the segments in descending order based on the remainders, and increment the number of actions for each segment by 1 according to the sorting order, until the number of actions is allocated.

[0182] Accordingly, the present invention also provides an electronic device including one or more processors, a memory, and one or more programs, wherein the one or more programs are stored in the memory and configured to be executed by the one or more processors, and the one or more programs include instructions for performing any of the methods.

[0183] The present invention also provides a readable storage medium, wherein the one or more programs include instructions that, when executed by a computing device, cause the computing device to perform any of the methods described in the method.

[0184] Those skilled in the art will understand that embodiments of this application can be provided as methods, systems, or computer program products. Therefore, this application can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, this application can take the form of a computer program product embodied on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0185] This application is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this application. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart... Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0186] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0187] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0188] The above are merely embodiments of the present invention and are not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention are included within the scope of the claims of the present invention pending approval.

Claims

1. A day-ahead discrete reactive power equipment action planning method based on predictive information, characterized in that, include: Based on the pre-determined main influencing factors of the target node voltage, obtain the day-ahead forecast information data corresponding to each factor; Input the day-ahead forecast information data corresponding to each factor into the voltage operation trend prediction model to obtain the day-ahead voltage operation trend prediction data; The current voltage operating trend forecast data includes several voltage forecast results; Based on the day-ahead voltage operating trend prediction data, day-ahead discrete reactive power equipment operation planning is performed, including: The day-ahead voltage trend forecast data is divided into a target number of segments using a trend segmentation method, with each segment having a duration of no less than a preset duration, including: If each voltage prediction result is taken as a segmentation point, the day-ahead voltage operating trend prediction data is divided into several segments. Calculate the voltage change between adjacent segments according to formula (1), find the two adjacent segments with the closest voltage change, merge the two segments, and at the same time, count the voltage fluctuation according to formula (2); repeat this step until the number of segments reaches the target number of segments and the duration of each segment is not less than the preset duration. （1） In the formula: Indicates the first Voltage prediction data for each segment point, This represents the voltage trend change between two adjacent segments with the closest change trends among N segments, where N represents the number of segments before each merging. （2） In the formula: Indicates the first Voltage fluctuation in each segment This represents the total voltage fluctuation. The number of actions of discrete reactive power equipment is allocated according to the voltage fluctuation of each segment, including: integer allocation of the number of actions of discrete reactive power equipment, with at least 1 action allocated to each segment, and the rest allocated according to the proportion of the voltage fluctuation of each segment to the total voltage fluctuation; if there are still action numbers remaining after integer allocation, the remaining action numbers are allocated according to the remainder.

2. The day-ahead discrete reactive power equipment action planning method based on predictive information according to claim 1, characterized in that, The main influencing factors of the target node voltage were determined according to the following method: Collect historical power grid operation data for the target area, including historical power grid operation data of the target node, its subordinate nodes, and adjacent nodes; Calculate the Pearson correlation coefficients between various data and the target node voltage; Data types with correlation coefficients exceeding a set value are selected as the main influencing factors of the target node voltage.

3. The day-ahead discrete reactive power equipment action planning method based on predictive information according to claim 2, characterized in that, The voltage operating trend prediction model is obtained through the following method: By combining the LSTM model with the Attention model, we obtain the Attention-LSTM model; Based on the main influencing factors of the target node voltage, relevant data for each factor are selected from the historical operation data of the power grid in the target area; The selected data for each factor are used as input to the Attention-LSTM model, and the target node voltage is used as output to train the Attention-LSTM model, thus obtaining a voltage trend prediction model.

4. The day-ahead discrete reactive power equipment action planning method based on predictive information according to claim 1, characterized in that, The integer allocation formula is as follows: ； In the formula: Indicates the first Number of actions for each segmented discrete quantity This indicates the total number of actions allowed per day. The integer represents the number of segments. Indicates the first Voltage fluctuation in each segment This represents the total voltage fluctuation.

5. The day-ahead discrete reactive power equipment action planning method based on predictive information according to claim 1, characterized in that, The process of allocating the remaining number of actions includes: Calculate the remainder of each segment after integer distribution using the following formula. : ； in, Indicates the first Voltage fluctuation in each segment This indicates the total number of actions allowed per day. For the number of segments, This represents the total voltage fluctuation, with % indicating the remainder. Sort the segments in descending order based on the remainders, and increment the number of actions for each segment by 1 according to the sorting order, until the number of actions is allocated.

6. A day-ahead discrete reactive power equipment action planning device based on predictive information, characterized in that, include: The acquisition module acquires the day-ahead forecast information data corresponding to each of the pre-determined main influencing factors of the target node voltage. The prediction module inputs the day-ahead prediction information data corresponding to each factor into the voltage operation trend prediction model to obtain day-ahead voltage operation trend prediction data; the day-ahead voltage operation trend prediction data includes several voltage prediction results; The action planning module performs day-ahead discrete reactive power equipment action planning based on the day-ahead voltage operating trend prediction data; The action planning module includes: The segmentation module uses the trend segmentation method to divide the day-ahead voltage operation trend prediction data into a target number of segments, and the duration of each segment is not less than the preset duration. The module includes: taking each voltage prediction result as a segmentation point, the day-ahead voltage operation trend prediction data is divided into several segments; calculating the voltage change between each adjacent segment according to formula (1), finding the two adjacent segments with the closest voltage change, merging the two segments, and repeating this step until the number of segments reaches the target number of segments and the duration of each segment is not less than the preset duration. （1） In the formula: Indicates the first Voltage prediction data for each segment point, This represents the voltage trend change between two adjacent segments with the closest change trends among N segments, where N represents the number of segments before each merging. The voltage fluctuation statistics module is used to find and merge two adjacent segments with the closest voltage changes in the segmentation module, and simultaneously calculate the voltage fluctuation according to formula (2): （2） In the formula: Indicates the first Voltage fluctuation in each segment This represents the total voltage fluctuation. The allocation module allocates the number of operations for discrete reactive power equipment based on the voltage fluctuation of each segment, including: The integer allocation module allocates the number of actions of discrete reactive power equipment to integers, with each segment allocated at least one action, and the rest allocated according to the proportion of the voltage fluctuation of each segment to the total voltage fluctuation. The remainder allocation module allocates the remaining action counts if there are still action counts remaining after integer allocation.

7. The day-ahead discrete reactive power equipment action planning device based on predictive information according to claim 6, characterized in that, It also includes: an influencing factor determination module, used for: Collect historical power grid operation data for the target area, including historical power grid operation data of the target node, its subordinate nodes, and adjacent nodes; Calculate the Pearson correlation coefficients between various data and the target node voltage; Data types with correlation coefficients exceeding a set value are selected as the main influencing factors of the target node voltage.

8. The day-ahead discrete reactive power equipment action planning device based on predictive information according to claim 7, characterized in that, The voltage operation trend prediction model adopts an Attention-LSTM model, and the device further includes a training module, which is used for: Based on the main influencing factors of the target node voltage, relevant data for each factor are selected from the historical operation data of the power grid in the target area; The selected data for each factor are used as input to the Attention-LSTM model, and the target node voltage is used as output to train the Attention-LSTM model, thus obtaining a voltage trend prediction model.

9. A day-ahead discrete reactive power equipment action planning device based on predictive information according to claim 6, characterized in that, The integer allocation module allocates according to the following formula: ； In the formula: Indicates the first Number of actions for each segmented discrete quantity This indicates the total number of actions allowed per day. The integer represents the number of segments. Indicates the first Voltage fluctuation in each segment This represents the total voltage fluctuation.

10. A day-ahead discrete reactive power equipment action planning device based on predictive information according to claim 6, characterized in that, The remainder allocation module is used for: Calculate the remainder of each segment after integer distribution using the following formula. : ； in, Indicates the first Voltage fluctuation in each segment This indicates the total number of actions allowed per day. For the number of segments, This represents the total voltage fluctuation, with % indicating the remainder. Sort the segments in descending order based on the remainders, and increment the number of actions for each segment by 1 according to the sorting order, until the number of actions is allocated.

11. An electronic device, characterized in that, It includes one or more processors, memory, and one or more programs, wherein the one or more programs are stored in the memory and configured to be executed by the one or more processors, and the one or more programs include instructions for performing any of the methods according to claims 1 to 5.

12. A readable storage medium, characterized in that, The one or more programs include instructions that, when executed by a computing device, cause the computing device to perform any of the methods according to claims 1 to 5.

Images