Electric power time sequence data fusion method for power grid planning optimization

Abstract

The invention relates to the technical field of data fusion, and provides a power time sequence data fusion method for power grid planning optimization, which comprises the following steps: acquiring a real-time load rate of a power grid, wind power and photovoltaic power of all power stations and meteorological data of all meteorological stations, and determining corrected meteorological data of the power stations, respectively establishing a wind power characteristic matrix and a photovoltaic characteristic matrix; constructing a dynamic adjustment adjacency matrix, determining a fusion power value according to the root-mean-square error of the wind power and the photovoltaic power of the power station, the real-time load rate of the power grid, and the wind power and the photovoltaic power of all the power stations, and determining the wind power and the photovoltaic power of the power station according to the fusion power value, the difference between the wind power and the photovoltaic power at adjacent acquisition moments. Determining a composite loss function; and obtaining a data fusion result according to the dynamic adjustment adjacency matrix, the composite loss function, the wind power characteristic matrix and the photovoltaic characteristic matrix. According to the invention, the accuracy of power time sequence data fusion of power grid planning optimization can be improved.

Description

Technical Field

[0001] The present application relates to the technical field of data fusion, and in particular to a method for fusion of electric power time series data for power grid planning optimization. Background Art

[0002] Power time-series data fusion for grid planning optimization integrates and processes power time-series data from different sources and time scales to improve the accuracy and optimization of grid planning. Data fusion improves data quality, reduces data redundancy, and provides more accurate and comprehensive basic data for grid planning, thereby enhancing the scientific nature and effectiveness of grid planning. Renewable energy is characterized by volatility and intermittency. Therefore, the supply of renewable energy from different sources is unstable, requiring the flexible adjustment of traditional thermal power on the load side and energy storage side. Therefore, renewable energy and thermal power have temporal and spatial complementarity, and renewable energy from different sources also have temporal and spatial complementarity.

[0003] When fusing power time series data, existing methods often ignore the complex spatiotemporal complementary relationship between wind power generation and photovoltaic power generation, lack constraints on the power upper limit, and lead to inaccurate power time series data fusion results for grid planning optimization. Summary of the Invention

[0004] This application provides a method for fusion of power time series data for grid planning optimization to solve the problem of ignoring the spatiotemporal complementarity between wind power generation and photovoltaic power generation, which leads to inaccurate power time series data fusion results for grid planning optimization. The technical solution adopted is as follows: An embodiment of the present application provides a method for fusion of power time series data for power grid planning optimization, the method comprising the following steps: Collect the real-time load rate of the power grid at different collection times, as well as the wind power and photovoltaic power of all power stations and the meteorological data of all weather stations. Determine the corrected meteorological data of the power station based on the distance between the power station and all weather stations and the meteorological data of all weather stations. Combine the wind power, photovoltaic power and location of all power stations at all collection times to establish the wind power characteristic matrix and photovoltaic characteristic matrix of the power station respectively. Based on the wind power and photovoltaic power of all power stations, a dynamically adjusted adjacency matrix is ​​constructed. The fused power value is determined based on the root mean square error of the wind power and photovoltaic power of the power station, the real-time load rate of the power grid, and the wind power and photovoltaic power of all power stations. The composite loss function is determined based on the fused power value, the difference between the wind power and the photovoltaic power at all adjacent acquisition moments. The data fusion result is obtained by dynamically adjusting the adjacency matrix, the composite loss function, the wind power characteristic matrix and the photovoltaic characteristic matrix of the power station.

[0005] Furthermore, the method of determining the corrected meteorological data of the power station based on the distance between the power station and all meteorological stations and the meteorological data of all meteorological stations includes the following specific methods: Meteorological data include wind speed and irradiance, and any kind of meteorological data is recorded as target meteorological data; The sum of the distances between the power station and all weather stations is recorded as the cumulative distance of the weather station to the power station, and the ratio of the distance between the power station and the weather station to the cumulative distance of the weather station to the power station is recorded as the distance weight between the power station and the weather station; The distance weight between the power station and the weather station is used as the weight of the target weather data of the weather station. The target weather data of all weather stations are weighted and summed, and the result of the weighted sum is recorded as the corrected target weather data of the power station.

[0006] Furthermore, the wind power characteristic matrix is ​​specifically: The wind power characteristic matrix contains three columns. The wind power of all power stations at the same collection time is arranged into one column to obtain the wind power column at the same collection time. The wind power of all collection times is arranged into one column in the order of the collection times to obtain the first column of the wind power characteristic matrix. Arrange the corrected wind speeds of all power stations at the same collection time into a column to obtain the corrected wind speed column at the same collection time; arrange the corrected wind speeds at all collection times into a column in the order of the collection times to obtain the second column of the wind power characteristic matrix; Arrange the locations of all power stations into a column to obtain a location column, and repeatedly fill the location column in the third column of the wind power characteristic matrix until the third column of the wind power characteristic matrix is ​​completely filled.

[0007] Furthermore, the photovoltaic characteristic matrix is ​​specifically: The photovoltaic characteristic matrix contains three columns. The photovoltaic power of all power stations at the same collection time is arranged into one column to obtain the photovoltaic power column at the same collection time. The photovoltaic power of all collection times is arranged into one column in the order of the collection times to obtain the first column of the photovoltaic characteristic matrix. Arranging the corrected irradiances of all power stations at the same collection time into a column to obtain the corrected irradiance column at the same collection time, arranging the corrected irradiances at all collection times into a column in the order of the collection times to obtain the second column of the photovoltaic characteristic matrix; Arrange the locations of all power stations into a column to obtain the location column, and repeatedly fill the location column in the third column of the photovoltaic characteristic matrix until the third column of the photovoltaic characteristic matrix is ​​completely filled.

[0008] Furthermore, the dynamic adjustment of the adjacency matrix is ​​specifically as follows: Based on all wind power of two different power stations, the similarity of wind power of two different power stations is determined; based on all photovoltaic power of two different power stations, the similarity of photovoltaic power of two different power stations is determined; Dynamically adjust the element values ​​corresponding to two different power stations in the adjacency matrix as follows: in, Indicates the dynamic adjustment of the power station in the adjacency matrix and power stations The corresponding element value; 、 、 represent respectively a preset first influence weight parameter, a second influence weight parameter, and a third influence weight parameter, the sum of which is 1; represents an exponential function with a natural constant as its base; Indicates a power station and power stations the distance between them; Indicates the preset attenuation parameter; Indicates a power station and power stations Similarity of wind power; Indicates a power station and power stations The photovoltaic power similarity.

[0009] Furthermore, the calculation formula of the fusion power value is: in, Indicates the fusion power value; represents the wind power weight; represents the photovoltaic power weight; Represents the cumulative sum of wind power from all power stations; Represents the cumulative sum of photovoltaic power of all power stations; Represents the weight adjustment factor.

[0010] Furthermore, the specific method for obtaining the wind power weight and the photovoltaic power weight is: The sum of the root mean square error of wind power and the number 1 is used as the denominator, and the number 1 is used as the value of the fraction of the numerator, which is recorded as the wind power weight; The sum of the root mean square error of the photovoltaic power and the number 1 is used as the denominator, and the number 1 is used as the value of the fraction of the numerator, which is recorded as the photovoltaic power weight.

[0011] Furthermore, the specific method for obtaining the weight adjustment factor is: The absolute value of the difference between the real-time load rate of the power grid and the preset economic security threshold of the power grid plus the number 1 is used as the denominator, and the value of the fraction with the number 1 as the numerator is recorded as the weight adjustment factor.

[0012] Furthermore, the composite loss function is specifically: Where, represents the composite loss function; 、 、 Respectively represent the preset first loss weight parameter, second loss weight parameter and third loss weight parameter, and the sum of the first loss weight parameter, the second loss weight parameter and the third loss weight parameter is 1; represents the mean square error of the ST-GCN model; Indicates the The absolute power difference at each acquisition moment, specifically, The collection time and The absolute value of the difference in wind power at the collection moment is recorded as The adjacent wind power difference at the collection time is The collection time and The absolute value of the difference in photovoltaic power at the time of collection is recorded as The adjacent photovoltaic power difference at the collection moment is The sum of the adjacent wind power difference and the adjacent photovoltaic power difference at the collection time is recorded as The absolute difference in power at each acquisition moment; Indicates the number of all acquisition moments; Indicates the preset safety load threshold of the power grid; Represents the maximum function.

[0013] Furthermore, the method of dynamically adjusting the adjacency matrix, the composite loss function, the wind power characteristic matrix and the photovoltaic characteristic matrix of the power station to obtain the data fusion result includes the following specific methods: The composite loss function is used as the loss function of the ST-GCN model, the dynamically adjusted adjacency matrix is ​​used as the dynamic adjacency matrix in the ST-GCN model, and the wind power feature matrix and photovoltaic feature matrix of the power station are input into the ST-GCN model to obtain the fusion result.

[0014] The beneficial effects of this application are: This application first takes into account that meteorological data directly affects the wind power generated by wind turbines, but there is a long distance between meteorological stations and power stations where wind turbines generate electricity. In order to improve the accuracy of the power fusion value obtained after the fusion of power time series data, accurate mapping is performed based on the meteorological data of each meteorological station to determine the accurate meteorological data of the location of each power station, that is, to determine the corrected meteorological data of the power station, and combine the wind power, photovoltaic power and position of all power stations at all collection times to establish the wind power characteristic matrix and photovoltaic characteristic matrix of the power station respectively; due to the high proportion of wind power generation and photovoltaic power generation accessed in the grid planning, wind power and photovoltaic power show the dual fluctuation characteristics of minute-level meteorological mutations and seasonal energy structure shifts, in order to capture the correlation between wind power and photovoltaic power of different power stations, as well as the impact of the locations of different power stations on their wind power and photovoltaic power, a dynamically adjusted adjacency matrix is ​​constructed; further, the wind power and photovoltaic power of the power station are adjusted. The influence weight of the photovoltaic power of the power station on the fused power value is determined. When one of the power data is affected, the fused power value is determined based on the other power data, thereby improving the accuracy of the fusion result. At the same time, the real-time load rate of the power grid is combined in the process of determining the fused power value, so that the real-time state of the power grid can be embedded in the data fusion process. The real-time carrying capacity of the power grid is considered in the data fusion process to avoid the problem of instantaneous overload of the power grid line caused by drastic fluctuations in wind power and photovoltaic power due to meteorological influences. Then, the composite loss function is determined to balance the prediction accuracy, power grid safety constraints and wind-solar synergy effects; finally, the data fusion results are obtained based on the dynamic adjustment of the adjacency matrix, the composite loss function, the wind power characteristic matrix and the photovoltaic characteristic matrix of the power station, so as to solve the problem of ignoring the spatiotemporal complementary effect between wind power generation and photovoltaic power generation, resulting in inaccurate results of the power time series data fusion for power grid planning optimization, thereby improving the accuracy of the power time series data fusion for power grid planning optimization. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative labor.

[0016] Figure 1 A flowchart of a method for fusion of electric power time series data for power grid planning optimization provided by one embodiment of the present application; Figure 2 A flowchart for obtaining corrected meteorological data provided in one embodiment of the present application. DETAILED DESCRIPTION

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

[0018] See also Figure 1 , which shows a flow chart of a method for fusion of power time series data for power grid planning optimization provided by one embodiment of the present application, the method comprising the following steps: Step S001: Collect the real-time load rate of the power grid at different collection times, as well as the wind power and photovoltaic power of all power stations and the meteorological data of all weather stations. Determine the corrected meteorological data of the power station based on the distance between the power station and all weather stations and the meteorological data of all weather stations. Combine the wind power, photovoltaic power and location of all power stations at all collection times to establish the wind power characteristic matrix and photovoltaic characteristic matrix of the power station respectively.

[0019] Wind power and photovoltaic power are collected from different power plants. Wind speed and irradiance are collected from different weather stations, and both wind speed and irradiance are recorded as meteorological data. The location of each weather station is collected using the GPS global positioning system. Smart meters are used to collect the real-time load rate of the power grid.

[0020] Preferably, in one embodiment of the present application, when collecting real-time load rate, wind power, photovoltaic power, and meteorological data, this embodiment collects real-time load rate, wind power, photovoltaic power, and meteorological data every 30 seconds, and collects real-time load rate, wind power, photovoltaic power, and meteorological data at a total of 100 collection moments. In actual application, as other implementation methods, implementers can determine the sampling frequency and number of real-time load rate, wind power, photovoltaic power, and meteorological data based on actual conditions, and this application does not impose any special restrictions.

[0021] In order to prevent missing data from affecting subsequent analysis, the missing real-time load rate, wind power, photovoltaic power, and meteorological data are filled. This embodiment selects the median filling method to fill the missing real-time load rate, wind power, photovoltaic power, and meteorological data. In actual application, as other implementation methods, on the basis of achieving the purpose of filling missing data, the implementer may use other existing methods such as median filling, mode filling, random sampling filling, K-nearest neighbor filling, regression filling, etc. to fill the missing data. This application does not impose any special restrictions.

[0022] It should be noted that, to facilitate computation, all wind power, photovoltaic power, and meteorological data involved in the calculations in this embodiment undergo data preprocessing to eliminate dimension effects. This embodiment uses the Z-Score standard normalization method to de-dimensionalize wind power, photovoltaic power, and meteorological data. In actual applications, implementers may use other methods, such as existing technologies such as maximum-minimum normalization, for de-dimensionalization, and this is not limited here.

[0023] Meteorological data directly affects the wind power generated by wind turbines. Wind turbine power stations are usually located in relatively open and spacious locations, while meteorological stations are often far away from the wind turbine power stations. In order to improve the accuracy of the power fusion value obtained after the fusion of power time series data, it is necessary to accurately map the meteorological data of each meteorological station and determine the accurate meteorological data of each power station location.

[0024] The cumulative distance between the power station and all weather stations is recorded as the power station's weather station cumulative distance, and the ratio of the distance between the power station and the weather station to the cumulative distance between the power station and the weather station is recorded as the distance weight between the power station and the weather station. Any type of weather data is recorded as the target weather data, and the distance weight between the power station and the weather station is used as the weight of the weather station's target weather data. The target weather data of all weather stations are weighted and summed, and the result of this weighted sum is recorded as the corrected target weather data for the power station.

[0025] The same method can be used to obtain the corrected meteorological data for any type of meteorological data. Figure 2 shown.

[0026] It can be understood that meteorological data includes wind speed and irradiance. Therefore, the corrected target meteorological data of the power station includes two types of data, wind speed and irradiance, so that the corrected wind speed and corrected irradiance of each power station can be obtained respectively; the corrected wind speed and corrected irradiance of the power station are the accurate wind speed and irradiance at the location of the power station respectively.

[0027] The wind power characteristic matrix and photovoltaic characteristic matrix of the power station are established according to the corrected wind speed, corrected irradiance, wind power, photovoltaic power and location of all power stations at all collection times.

[0028] Preferably, as an embodiment of the present application, the wind power characteristic matrix contains a total of three columns, the wind power of all power stations at the same collection time is arranged into one column, the wind power column at the same collection time is obtained, the wind power at all collection times is arranged into one column in the order of collection time, and the first column of the wind power characteristic matrix is ​​obtained; the corrected wind speed of all power stations at the same collection time is arranged into one column, the corrected wind speed column at the same collection time is obtained, the corrected wind speed at all collection times is arranged into one column in the order of collection time, and the second column of the wind power characteristic matrix is ​​obtained; the positions of all power stations are arranged into one column, the position column is obtained, and the position column is filled in the third column of the wind power characteristic matrix in sequence until there is no vacant position in the third column of the wind power characteristic matrix.

[0029] Preferably, as an embodiment of the present application, the photovoltaic characteristic matrix contains three columns in total, the photovoltaic powers of all power stations at the same collection time are arranged into one column, the photovoltaic power column at the same collection time is obtained, the photovoltaic powers at all collection times are arranged into one column in the order of collection time, and the first column of the photovoltaic characteristic matrix is ​​obtained; the corrected irradiances of all power stations at the same collection time are arranged into one column, the corrected irradiance column at the same collection time is obtained, the corrected irradiances at all collection times are arranged into one column in the order of collection time, and the second column of the photovoltaic characteristic matrix is ​​obtained; the positions of all power stations are arranged into one column, the position column is obtained, and the position column is filled in the third column of the photovoltaic characteristic matrix in sequence until there is no vacant position in the third column of the photovoltaic characteristic matrix.

[0030] The same row of the wind power characteristic matrix and the photovoltaic characteristic matrix of the power station corresponds to the same power station, and the arrangement order of the power stations corresponding to the data in each column is the same.

[0031] At this point, the real-time load rate of the power grid at different collection times, as well as the wind power, photovoltaic power, wind power characteristic matrix and photovoltaic characteristic matrix of all power stations, and meteorological data and corrected meteorological data of all weather stations are obtained.

[0032] Step S002: Based on the wind power and photovoltaic power of all power stations, a dynamically adjusted adjacency matrix is ​​constructed; based on the root mean square error of the wind power and photovoltaic power of the power station, the real-time load rate of the power grid, and the wind power and photovoltaic power of all power stations, a fused power value is determined; based on the fused power value, the difference between the wind power at all adjacent acquisition moments, and the difference between the photovoltaic power, a composite loss function is determined.

[0033] Topographic disturbances in wind speed produce high-frequency fluctuations, and wind turbine power response lags behind wind speed changes. Irradiance, affected by cloud movement, produces minute-by-minute power dips, and the power curve exhibits a strong periodic correlation with solar altitude. Consequently, in grid planning with a high proportion of wind and photovoltaic power generation, wind and photovoltaic power generation exhibit dual fluctuations characterized by sudden minute-by-minute meteorological changes and seasonal shifts in the energy mix.

[0034] The ST-GCN (Spatial-Temporal Graph Convolutional Networks) model is a deep learning model that combines the GCN graph convolutional network and the TCN temporal convolutional network, combining information processing capabilities in both spatial and temporal dimensions. The ST-GCN model captures the spatial dependencies between nodes in a graph structure through a graph convolution layer, while simultaneously capturing the temporal dependencies in time series data through a temporal convolution layer. This effectively captures complex relationships in spatiotemporal data. Therefore, this application uses the ST-GCN model to implement power time series data fusion for grid planning optimization.

[0035] Furthermore, wind power and photovoltaic power exhibit dynamic temporal variations and spatial geographic correlations. Wind speed fluctuates minute-by-minute due to turbulence. Adjacent wind farms have similar meteorological conditions, such as wind speed and cloud cover. Consequently, wind power and photovoltaic power at adjacent wind farms exhibit a clear correlation. PV power clusters are influenced by the diurnal cycle, with photovoltaic power high during the day and zero at night. Influenced by topography and light distribution, irradiance is evenly distributed across different regions, and power fluctuates synchronously.

[0036] In order to capture the correlation between wind power and photovoltaic power of different power stations, as well as the impact of the location of different power stations on their wind power and photovoltaic power, a dynamically adjusted adjacency matrix is ​​constructed.

[0037] Arrange all wind power of the power station in the order of the wind power collection time to obtain the wind power sequence of the power station, and record the Pearson correlation coefficient between the wind power sequences of two different power stations as the wind power similarity of the two different power stations; arrange all photovoltaic power of the power station in the order of the photovoltaic power collection time to obtain the photovoltaic power sequence of the power station, and record the Pearson correlation coefficient between the photovoltaic power sequences of two different power stations as the photovoltaic power similarity of the two different power stations.

[0038] Preferably, as an embodiment of the present application, the element values ​​corresponding to two different power stations in the adjacency matrix are dynamically adjusted as follows: in, Indicates the dynamic adjustment of the power station in the adjacency matrix and power stations The corresponding element value; 、 、 Respectively represent the preset first influence weight parameter, the second influence weight parameter and the third influence weight parameter, the sum of the first influence weight parameter, the second influence weight parameter and the third influence weight parameter is 1 and they are all preset parameter values. In this embodiment, the values ​​of the first influence weight parameter, the second influence weight parameter and the third influence weight parameter are all ; represents an exponential function with a natural constant as its base; Indicates a power station and power stations the distance between them; represents the preset attenuation parameter. In this embodiment, the attenuation parameter value is 15; Indicates a power station and power stations Similarity of wind power; Indicates a power station and power stations The photovoltaic power similarity.

[0039] The ST-GCN model can obtain the root mean square error of wind power and photovoltaic power of the power station respectively.

[0040] Among them, the calculation of the root mean square error of the ST-GCN model is a well-known technology and will not be repeated here. Specifically, the ST-GCN model can obtain the wind power prediction value and photovoltaic power prediction value of the power station. The cumulative sum of the absolute values ​​of the differences between the wind power prediction values ​​and the wind power of all power stations is the root mean square error of the wind power of the power station; the cumulative sum of the absolute values ​​of the differences between the photovoltaic power prediction values ​​and the photovoltaic power of all power stations is the root mean square error of the photovoltaic power of the power station.

[0041] The fusion power value is determined based on the root mean square error of the wind power and photovoltaic power of the power station, the real-time load rate of the power grid, and the wind power and photovoltaic power of all power stations.

[0042] in, Indicates the fusion power value; represents the wind power weight; represents the photovoltaic power weight; represents the root mean square error of wind power; represents the root mean square error of photovoltaic power; represents the weight adjustment factor; Indicates the real-time load rate of the power grid; Indicates the economic security threshold of the power grid, which is 85% of the maximum load of the power grid; Represents the cumulative sum of wind power from all power stations; Represents the cumulative sum of the photovoltaic power of all power plants.

[0043] Based on the wind power weight and photovoltaic power weight determined by the root mean square error of the wind power and photovoltaic power of each power station, the weights of the impact of the "cumulative sum of wind power from all power stations" and the "cumulative sum of photovoltaic power from all power stations" on the fused power value can be adjusted. When wind power is more significantly affected by complex meteorological factors such as wind speed and turbulence, the root mean square error of wind power is larger and the wind power weight is smaller. In this case, the cumulative sum of photovoltaic power from all power stations has a greater impact on the fused power value, which can improve the accuracy of the fusion result. When photovoltaic power is more significantly affected by cloud cover, the root mean square error of photovoltaic power is larger and the photovoltaic power weight is smaller. In this case, the cumulative sum of wind power from all power stations has a greater impact on the fused power value, which can improve the accuracy of the fusion result.

[0044] The weight adjustment factor determined according to the real-time load rate of the power grid can embed the real-time status of the power grid into the data fusion process. The real-time carrying capacity of the power grid is taken into account during the data fusion process to avoid the problem of instantaneous overload of the power grid lines caused by drastic fluctuations in wind power and photovoltaic power due to meteorological influences.

[0045] The closer the grid's real-time load rate is to its economic safety threshold, and the lower the real-time load rate is, the greater the renewable energy penetration rate. Furthermore, when the real-time load rate is higher than the economic safety threshold, grid overload can be automatically avoided.

[0046] The weight of the fused power value is dynamically adjusted according to the root mean square error of the wind power and photovoltaic power of the power station and the real-time load rate of the power grid, as well as the process of determining the composite loss function. This can balance the prediction accuracy, power grid security constraints and the synergistic effect of wind and solar power. At the same time, it can adapt to sudden meteorological changes such as sudden drop in wind speed and cloud cover. The smaller the root mean square error of the ST-GCN model, the more the fused power value should dominate the fusion result, and the higher the credibility of the fusion value obtained by the ST-GCN model.

[0047] Furthermore, the three goals of jointly optimizing prediction accuracy, optimizing power smoothness and optimizing grid security are constructed to construct a composite loss function to ensure the accuracy of the fusion results while meeting the actual engineering constraints.

[0048] Specifically, the composite loss function is determined based on the fusion power value, the difference between the wind power and the photovoltaic power at all adjacent acquisition moments. .

[0049] Preferably, as an embodiment of the present application, the composite loss function Specifically: Where, represents the composite loss function; 、 、 Respectively represent the preset first loss weight parameter, second loss weight parameter and third loss weight parameter, the sum of the first loss weight parameter, the second loss weight parameter and the third loss weight parameter is 1 and they are all preset parameter values. In this embodiment, the values ​​of the first loss weight parameter, the second loss weight parameter and the third loss weight parameter are all ; represents the mean square error of the ST-GCN model; Indicates the The absolute power difference at each acquisition moment, specifically, The collection time and The absolute value of the difference in wind power at the collection moment is recorded as The adjacent wind power difference at the collection time is The collection time and The absolute value of the difference in photovoltaic power at the time of collection is recorded as The adjacent photovoltaic power difference at the collection moment is The sum of the adjacent wind power difference and the adjacent photovoltaic power difference at the collection time is recorded as The absolute difference in power at each acquisition moment; Indicates the number of all acquisition moments; Indicates the safe load threshold of the power grid, which is 90% of the maximum load of the power grid; Indicates the fusion power value; Represents the maximum function. Specifically, the maximum function takes the maximum value among the values ​​separated by commas in brackets.

[0050] Among them, the calculation of the mean square error of the ST-GCN model is a well-known technology and will not be repeated here. Specifically, the ST-GCN model can obtain the wind power prediction value and photovoltaic power prediction value of the power station. The square of the difference between the wind power prediction value and the wind power of the power station is recorded as the first square, and the square of the difference between the photovoltaic power prediction value and the photovoltaic power of the power station is recorded as the second square. The sum of the first square and the second square is the mean square error of the ST-GCN model.

[0051] Wind power and photovoltaic power at power stations are significantly affected by meteorological fluctuations. High-precision forecasting can reduce grid dispatch errors. In the composite loss function, mean square error (MSE) is used to mitigate the impact of extreme errors on the grid. Smoothing constraints can also reduce the pressure on grid frequency regulation caused by fluctuations in wind and photovoltaic power. The difference between wind and photovoltaic power at adjacent acquisition times is more sensitive to small fluctuations, encouraging a smooth transition in the composite loss function. The grid's carrying capacity is the most critical constraint, helping the model approach safety boundaries while strictly restricting out-of-bounds behavior.

[0052] At this point, the dynamically adjusted adjacency matrix and composite loss function are obtained.

[0053] Step S003: Obtain data fusion results based on the dynamically adjusted adjacency matrix, the composite loss function, the wind power characteristic matrix and the photovoltaic characteristic matrix of the power station.

[0054] The composite loss function is used as the loss function of the ST-GCN model, the dynamically adjusted adjacency matrix is ​​used as the dynamic adjacency matrix in the ST-GCN model, and the wind power feature matrix and photovoltaic feature matrix of the power station are input into the ST-GCN model. According to the dynamically adjusted adjacency matrix and loss function, the ST-GCN model can output the fusion result.

[0055] Among them, using the ST-GCN model to obtain data fusion results is a well-known technology and will not be described in detail.

[0056] At this point, the fusion of power time series data for power grid planning optimization has been achieved.

[0057] The above description is only a preferred embodiment of the present application and is not intended to limit the present application. Any modifications, equivalent replacements, improvements, etc. made within the principles of the present application shall be included in the scope of protection of the present application.

Claims

1. A method for fusion of power time series data for power grid planning optimization, characterized in that: The method comprises the following steps: Collect the real-time load rate of the power grid at different collection times, as well as the wind power and photovoltaic power of all power stations and the meteorological data of all weather stations. Determine the corrected meteorological data of the power station based on the distance between the power station and all weather stations and the meteorological data of all weather stations. Combine the wind power, photovoltaic power and location of all power stations at all collection times to establish the wind power characteristic matrix and photovoltaic characteristic matrix of the power station respectively. Based on the wind power and photovoltaic power of all power stations, a dynamically adjusted adjacency matrix is ​​constructed. The fused power value is determined based on the root mean square error of the wind power and photovoltaic power of the power station, the real-time load rate of the power grid, and the wind power and photovoltaic power of all power stations. The composite loss function is determined based on the fused power value, the difference between the wind power and the photovoltaic power at all adjacent acquisition moments. The data fusion result is obtained by dynamically adjusting the adjacency matrix, the composite loss function, the wind power characteristic matrix and the photovoltaic characteristic matrix of the power station.

2. The method for fusion of electric power time series data for power grid planning optimization according to claim 1, characterized in that: The method of determining the corrected meteorological data of the power station based on the distance between the power station and all meteorological stations and the meteorological data of all meteorological stations includes: Meteorological data include wind speed and irradiance, and any kind of meteorological data is recorded as target meteorological data; The sum of the distances between the power station and all weather stations is recorded as the cumulative distance of the weather station to the power station, and the ratio of the distance between the power station and the weather station to the cumulative distance of the weather station to the power station is recorded as the distance weight between the power station and the weather station; The distance weight between the power station and the weather station is used as the weight of the target weather data of the weather station. The target weather data of all weather stations are weighted and summed, and the result of the weighted sum is recorded as the corrected target weather data of the power station.

3. The method for fusion of electric power time series data for power grid planning optimization according to claim 2, characterized in that: The wind power characteristic matrix is ​​specifically: The wind power characteristic matrix contains three columns. The wind power of all power stations at the same collection time is arranged into one column to obtain the wind power column at the same collection time. The wind power of all collection times is arranged into one column in the order of the collection times to obtain the first column of the wind power characteristic matrix. Arrange the corrected wind speeds of all power stations at the same collection time into a column to obtain the corrected wind speed column at the same collection time; arrange the corrected wind speeds at all collection times into a column in the order of the collection times to obtain the second column of the wind power characteristic matrix; Arrange the locations of all power stations into a column to obtain a location column, and repeatedly fill the location column in the third column of the wind power characteristic matrix until the third column of the wind power characteristic matrix is ​​completely filled.

4. The method for fusion of electric power time series data for power grid planning optimization according to claim 2, characterized in that: The photovoltaic characteristic matrix is ​​specifically: The photovoltaic characteristic matrix contains three columns. The photovoltaic power of all power stations at the same collection time is arranged into one column to obtain the photovoltaic power column at the same collection time. The photovoltaic power of all collection times is arranged into one column in the order of the collection times to obtain the first column of the photovoltaic characteristic matrix. Arranging the corrected irradiances of all power stations at the same collection time into a column to obtain the corrected irradiance column at the same collection time, arranging the corrected irradiances at all collection times into a column in the order of the collection times to obtain the second column of the photovoltaic characteristic matrix; Arrange the locations of all power stations into a column to obtain the location column, and repeatedly fill the location column in the third column of the photovoltaic characteristic matrix until the third column of the photovoltaic characteristic matrix is ​​completely filled.

5. The method for fusion of electric power time series data for power grid planning optimization according to claim 1, characterized in that: The specific steps of dynamically adjusting the adjacency matrix are as follows: Based on all wind power of two different power stations, the similarity of wind power of two different power stations is determined; based on all photovoltaic power of two different power stations, the similarity of photovoltaic power of two different power stations is determined; Dynamically adjust the element values ​​corresponding to two different power stations in the adjacency matrix as follows: in, Indicates the dynamic adjustment of the power station in the adjacency matrix and power stations The corresponding element value; 、 、 represent respectively a preset first influence weight parameter, a second influence weight parameter, and a third influence weight parameter, the sum of which is 1; represents an exponential function with a natural constant as its base; Indicates a power station and power stations the distance between them; Indicates the preset attenuation parameter; Indicates a power station and power stations Similarity of wind power; Indicates a power station and power stations The photovoltaic power similarity.

6. The method for fusion of electric power time series data for power grid planning optimization according to claim 1, characterized in that: The calculation formula of the fusion power value is: in, Indicates the fusion power value; represents the wind power weight; represents the photovoltaic power weight; Represents the cumulative sum of wind power from all power stations; Represents the cumulative sum of photovoltaic power of all power stations; Represents the weight adjustment factor.

7. The method for fusion of electric power time series data for power grid planning optimization according to claim 6, characterized in that: The specific method for obtaining the wind power weight and the photovoltaic power weight is: The sum of the root mean square error of wind power and the number 1 is used as the denominator, and the number 1 is used as the value of the fraction of the numerator, which is recorded as the wind power weight; The sum of the root mean square error of the photovoltaic power and the number 1 is used as the denominator, and the number 1 is used as the value of the fraction of the numerator, which is recorded as the photovoltaic power weight.

8. The method for fusion of electric power time series data for power grid planning optimization according to claim 6, characterized in that: The specific method for obtaining the weight adjustment factor is: The absolute value of the difference between the real-time load rate of the power grid and the preset economic security threshold of the power grid plus the number 1 is used as the denominator, and the value of the fraction with the number 1 as the numerator is recorded as the weight adjustment factor.

9. The method for fusion of electric power time series data for power grid planning optimization according to claim 6, characterized in that: The composite loss function is specifically: Where, represents the composite loss function; 、 、 Respectively represent the preset first loss weight parameter, second loss weight parameter and third loss weight parameter, and the sum of the first loss weight parameter, the second loss weight parameter and the third loss weight parameter is 1; represents the mean square error of the ST-GCN model; Indicates the The absolute power difference at each acquisition moment, specifically, The collection time and The absolute value of the difference in wind power at the collection moment is recorded as The adjacent wind power difference at the collection time is The collection time and The absolute value of the difference in photovoltaic power at the time of collection is recorded as The adjacent photovoltaic power difference at the collection moment is The sum of the adjacent wind power difference and the adjacent photovoltaic power difference at the collection time is recorded as The absolute difference in power at each acquisition moment; Indicates the number of all acquisition moments; Indicates the preset safety load threshold of the power grid; Represents the maximum function.

10. The method for fusion of electric power time series data for power grid planning optimization according to claim 1, characterized in that: The method of dynamically adjusting the adjacency matrix, the composite loss function, the wind power characteristic matrix and the photovoltaic characteristic matrix of the power station to obtain the data fusion result includes the following specific methods: The composite loss function is used as the loss function of the ST-GCN model, the dynamically adjusted adjacency matrix is ​​used as the dynamic adjacency matrix in the ST-GCN model, and the wind power feature matrix and photovoltaic feature matrix of the power station are input into the ST-GCN model to obtain the fusion result.

Images