A power meteorological prediction error analysis method based on transmission line attribute weighting

Abstract

The application discloses a power meteorological prediction error analysis method based on transmission line attribute weighting, obtains geographical information data and voltage grade attribute data of a transmission line in a target area; obtains meteorological grid point prediction data and live data in the same historical period and in the same geographical range; for each meteorological grid point, the spatial distance weight, the voltage grade weight and the tower density weight of the meteorological grid point relative to each adjacent transmission line are calculated, the comprehensive weight coefficient of the meteorological grid point is calculated, and a weight coefficient matrix is formed; the error value between the prediction data and the live data of each meteorological grid point is calculated; the error value of each meteorological grid point is weighted by using the weight coefficient matrix, and a weighted area error index reflecting the meteorological prediction accuracy around the transmission line is calculated. Through the comprehensive weight coefficient, the meteorological prediction error evaluation system facing the safety of the power system is constructed, and a scientific basis can be provided for micro-meteorological device site selection.

Description

Technical Field

[0001] This invention belongs to the field of meteorological data processing technology, and relates to a method for analyzing power meteorological forecasting errors based on power transmission line attribute weighting. Background Technology

[0002] Currently, power weather forecasting typically employs numerical weather prediction (NWP) models, whose output is uniform grid data. Error analysis usually uses metrics such as root mean square error (RMSE) and mean absolute error (MAE), treating all grid points equally. Existing error analysis methods neglect the importance of transmission lines. Even the most accurate prediction at a grid point far from a transmission line has no direct impact on grid security; conversely, a prediction error at a grid point near a transmission line can directly lead to operational and maintenance decision-making errors. Furthermore, the importance of lines varies across different voltage levels, resulting in vastly different social and economic impacts from their outages. Existing error analysis fails to reflect these differences, preventing optimization resources from being concentrated on the most critical lines. Therefore, there is an urgent need for a differentiated error analysis method that reflects the characteristics of the power system. Summary of the Invention

[0003] The purpose of this invention is to provide a power meteorological forecasting error analysis method based on transmission line attributes. By using spatial distance weight, voltage level weight, and tower density weight to obtain a comprehensive weight coefficient, a meteorological forecasting error assessment system for power system safety is constructed using the comprehensive weight coefficient, which is helpful for the training and optimization of meteorological forecasting models.

[0004] To achieve the above objectives, the technical solution proposed in this invention is: a method for analyzing power meteorological forecasting errors based on transmission line attribute weighting, comprising the following steps:

[0005] S1: Obtain geographic information data and voltage level attribute data of transmission lines within the target area; obtain meteorological grid prediction data and real-time data within the same historical period and geographic area;

[0006] S2: For each meteorological grid point, calculate the spatial distance weight, voltage level weight, and tower density weight of the meteorological grid point relative to each adjacent transmission line, calculate the temporary comprehensive weight coefficient of the meteorological grid point, and take the maximum value as the final comprehensive weight coefficient of the meteorological grid point. The final comprehensive weight coefficients of all meteorological grid points constitute the weight coefficient matrix.

[0007] S3: Calculate the error between the predicted data and the actual data at each meteorological grid point;

[0008] S4: The error value of each meteorological grid point is weighted using the weighting coefficient matrix to calculate the weighted regional error index that reflects the accuracy of meteorological forecasts around the transmission line.

[0009] Specifically, the temporary comprehensive weighting coefficient of the meteorological grid point is the product of the spatial distance weight, voltage level weight, and tower density weight of the meteorological grid point relative to each adjacent transmission line.

[0010] Specifically, the spatial distance weight The calculation method is as follows: Calculate the shortest spatial distance from the meteorological grid point to all transmission lines. Spatial distance weights are calculated using an exponential decay function or a Gaussian function. ;

[0011] ;

[0012] Where L is the characteristic attenuation distance. This is the spatial distance weighting correction coefficient.

[0013] Specifically, find the meteorological grid points Find the nearest transmission line and obtain its voltage level V. Then, retrieve its voltage level weight by querying a predefined voltage level-weight mapping table. .

[0014] Specifically, the voltage level weights are adjusted according to the voltage level weight correction factor:

[0015] ;

[0016] in This is a voltage level weighting correction factor; In order to be consistent with meteorological grids The voltage of the nearest transmission line; This is a voltage level-weight mapping function. It is a natural constant.

[0017] Specifically, based on the current meteorological grid Centered on a circular statistical area with a radius of K kilometers, the number of all transmission line towers within this area is counted. The tower density weight is calculated using the following formula. :

[0018] ;

[0019] in, This is the tower density weighting correction coefficient.

[0020] Specifically, step S3 involves calculating the value of each meteorological grid point. Mean absolute error Introducing error analysis weighting factors The mean absolute error is corrected, and calculations are performed for each meteorological grid point. Overall prediction error:

[0021] .

[0022] Specifically, the error analysis weighting factors are obtained as follows:

[0023] Each meteorological grid point is on the [number]th The predicted and actual event values ​​at each time step are converted into binary data based on event thresholds, where 1 indicates the event occurred and 0 indicates the event did not occur. Thus, for the entire analysis area and time period, a binary field of the predicted and actual event values ​​is obtained:

[0024] ;

[0025] Where TS is the event forecast score, H is the number of meteorological grid points that were predicted to occur and actually occurred, FA is the number of meteorological grid points that were predicted to occur but did not actually occur, and M is the number of meteorological grid points that were missed but did not occur.

[0026] The calculated event forecast score (TS score) is converted into error analysis weighting factors. The conversion rules are as follows:

[0027] ;

[0028] in, It is a positive number, used to prevent the denominator from being zero.

[0029] Specifically, in step S4, the prediction comprehensive error is first multiplied by the final comprehensive weight coefficient to calculate the weighted error of each meteorological grid point, and then the weighted average absolute error of the entire region is calculated based on the weighted error of each meteorological grid point.

[0030] The present invention provides a computer-readable storage medium having a computer program stored thereon, wherein the computer program, when executed by a processor, implements the power meteorological forecast error analysis method.

[0031] The present invention provides an electronic device, including a memory, a processor, and a computer program stored in the memory, wherein the processor executes the computer program to implement the power meteorological forecast error analysis method.

[0032] Compared with the prior art, the advantages of this application are:

[0033] 1. This application constructs a meteorological forecasting error assessment system for power system security by introducing a multi-weighting mechanism of spatial distance weight, voltage level weight, and tower density weight. It accurately focuses limited model optimization computing resources on the core area and key assets of the transmission corridor, thereby improving the directional performance of the meteorological forecasting model in power-sensitive areas and providing more reliable and targeted high-precision meteorological data support for power grid disaster prevention and mitigation decisions.

[0034] 2. The comprehensive weight coefficient calculation method proposed in this application has high flexibility and scalability. The weight factors can be dynamically adjusted according to different geographical environments, climate characteristics and operation and maintenance strategies. The weight correction coefficient can be set according to the disaster-causing mechanism of different power meteorological disasters. For forecast error analysis and improvement, it can truly reflect the engineering practical value of the forecast data.

[0035] 3. The weighted regional error index obtained in this application can be combined with the meteorological forecast model parameter optimization process to form an automated and intelligent closed-loop feedback improvement system, which guides the optimization of meteorological forecast model parameters by using the weighted regional error index as a loss function. Attached Figure Description

[0036] Figure 1 This is a flowchart of the method of the present invention. Detailed Implementation

[0037] To make the objectives, technical solutions, and advantages of the present invention clearer, the specific embodiments of the present invention will be described in detail below.

[0038] refer to Figure 1 In this embodiment, a method for analyzing power meteorological forecasting errors based on transmission line attributes is provided, including the following steps:

[0039] S1: Obtain geographic information data and voltage level attribute data of transmission lines within the target area; obtain meteorological grid prediction data and real-time data within the same historical period and geographic area;

[0040] S2: For each meteorological grid point, calculate the spatial distance weight, voltage level weight, and tower density weight of the meteorological grid point relative to each adjacent transmission line, calculate the temporary comprehensive weight coefficient of the meteorological grid point, and take the maximum value as the final comprehensive weight coefficient of the meteorological grid point. The final comprehensive weight coefficients of all meteorological grid points constitute the weight coefficient matrix.

[0041] S3: Calculate the error between the predicted data and the actual data at each meteorological grid point;

[0042] S4: The error value of each meteorological grid point is weighted using the weighting coefficient matrix to calculate the weighted regional error index that reflects the accuracy of meteorological forecasts around the transmission line.

[0043] In this embodiment, step S1 specifically involves extracting the latitude and longitude coordinates of all transmission lines from the power grid GIS system, as well as the voltage level (e.g., 500kV, 220kV, 110kV, 60kV, etc.) corresponding to each line.

[0044] Predictive data of meteorological grid points are obtained from meteorological forecasting models or other channels. Real-time data of basic meteorological elements for the same period are obtained from meteorological departments or meteorological stations and micro-meteorological stations deployed along power transmission lines. The real-time data needs to be preprocessed and interpolated or aggregated onto the same meteorological grid points as the forecast data to form real-time meteorological grid point data with the same temporal and spatial resolution.

[0045] In this embodiment, step S2 specifically includes:

[0046] For each meteorological grid point , For the row coordinates of the meteorological grid, To assign coordinates to meteorological grid points, perform the following operations:

[0047] A. Spatial distance weighting Calculate: Calculate the shortest spatial distance from this meteorological grid point to all power transmission lines. Spatial distance weights are calculated using an exponential decay function or a Gaussian function. ;

[0048] ;

[0049] Where L is the characteristic attenuation distance, which can be set according to the spatial resolution of the actual prediction data (50 km is recommended). This is the spatial distance weighting correction coefficient, typically 1, but can be set according to the specific application scenario. The closer the meteorological grid is to the transmission line, the closer the spatial distance weight is to 1 (most important); the farther the meteorological grid is from the transmission line, the closer the spatial distance weight is to 0.

[0050] B. Voltage Level Weighting Calculation: Find the meteorological grid points Find the nearest transmission line and obtain its voltage level V. Then, retrieve its voltage level weight by querying a predefined voltage level-weight mapping table. The voltage level-weight mapping table is shown in Table 1. The rule for setting it is that transmission lines with higher voltage levels have a wider impact range and enjoy higher voltage level weights.

[0051] Table 1 Voltage Level Weighting Table

[0052]

[0053] Depending on the importance of power grid security, the voltage level-weight mapping table may have some discrepancies, which can be corrected using a voltage level weighting correction factor. The adjustment is made, and the calculation formula is as follows:

[0054] ;

[0055] in This is the voltage level weighting correction factor, which is generally 1, depending on the application. In order to be consistent with meteorological grids The voltage of the nearest transmission line; Voltage level - weight, It is a natural constant.

[0056] C. Tower density weight Calculation: Statistically determine the pole density in the area surrounding the meteorological grid point. The higher the pole density, the greater the weight of the pole density.

[0057] Based on current meteorological grid A circular statistical area with a radius of K kilometers is defined centered on the target (here, the reference value for K is set to 10 kilometers, which needs to be adjusted according to the predicted meteorological grid resolution). The number of all transmission line towers within this area is then counted. The tower density weight is calculated using the following formula. :

[0058] ;

[0059] in, This is the tower density weighting correction coefficient, typically set to 0.1, used to adjust the magnitude of the influence of tower density. The logarithmic function prevents excessive weighting in areas with extremely dense tower density, ensuring a smooth change in tower density weighting.

[0060] D. Calculate the spatial distance weight, voltage level weight, and tower density weight of the meteorological grid point relative to each adjacent transmission line, and calculate the temporary comprehensive weight coefficient of the meteorological grid point;

[0061] ;

[0062] E. Select the maximum value from all calculated temporary comprehensive weight coefficients as the final comprehensive weight coefficient for that meteorological grid point; the final comprehensive weight coefficients of all meteorological grid points constitute the weight coefficient matrix.

[0063] In this embodiment, step S3 specifically includes:

[0064] Calculate each meteorological grid point The prediction error can be selected as the mean absolute error. The root mean square error (RMSE) and the mean error (ME) are used as prediction errors. Here, the mean absolute error is selected as the prediction error.

[0065] ;

[0066] in, It is a meteorological grid ( In the The predicted value at each time step, It is a meteorological grid ( In the The actual value at each time step, For time step index, This represents the total number of time steps.

[0067] To further focus on high-impact weather events, an error analysis weighting factor can be introduced in step S3. The mean absolute error is corrected, and calculations are performed for each meteorological grid point. The overall prediction error :

[0068] ;

[0069] Events such as heavy rain (hourly rainfall ≥ 20 mm) and strong winds (wind speed ≥ level 10) are included. The event thresholds are set according to the needs of model evaluation. Their impact on the power grid is more severe than that of normal weather and requires more attention. Therefore, they are included in the error calculation, and their event forecast score TS (Threat Score), also known as the Critical Success Index (CSI), is calculated.

[0070] The specific calculation details are as follows: for each meteorological grid point (i, j) at the th... The predicted and actual event values ​​at each time step are converted into binary data (1 indicates the event occurred, 0 indicates the event did not occur) based on the event threshold. Thus, for the entire analysis area and time period, a binary field of the predicted and actual event values ​​can be obtained.

[0071] ;

[0072] Where H represents the number of meteorological grid points that were predicted to occur and actually occurred, FA represents the number of meteorological grid points that were predicted to occur but did not actually occur, and M represents the number of meteorological grid points that were missed but actually occurred.

[0073] The calculated event forecast score (TS score) is converted into error analysis weighting factors. The conversion rules are as follows:

[0074] ;

[0075] in, The value should be a very small positive number (e.g., 0.01) to prevent the denominator from being zero. The better the forecast of high-impact weather on the power grid, the higher the event forecast score (TS), and the higher the error analysis weighting factor. The smaller the value, for example, when TS=0.8, ≈1.23); the worse the forecast of high-impact weather on the power grid, the lower the event forecast score TS, and the greater the error analysis weighting factor. The larger the value, the more it guides the model's performance. This cleverly achieves the guiding effect that the worse the model's forecasting ability for key weather events, the greater the amplification of its error in the overall assessment.

[0076] For meteorological elements for which no meteorological events are specified, the event forecast score (TS) is typically not calculated. In this case, the default error analysis weighting factor is used. = 1.0 indicates that no additional weighting is applied.

[0077] In this embodiment, S4 specifically refers to:

[0078] The weighted error for each meteorological grid point is:

[0079] ;

[0080] Calculate the weighted average absolute error over the entire domain. :

[0081] ;

[0082] Traditional methods calculate the simple average (MAE) of errors at all meteorological grid points, while WMAE better reflects the error level in areas with a greater impact on power grid security.

[0083] Another embodiment of the present invention provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the power meteorological forecasting error analysis method.

[0084] Another embodiment of the present invention provides an electronic device, including a memory, a processor, and a computer program stored in the memory, wherein the processor executes the computer program to implement the power meteorological forecast error analysis method.

[0085] The above description merely illustrates preferred embodiments of the present invention and is not intended to limit the invention in any other way. Any person skilled in the art may make modifications based on the above disclosure to obtain equivalent embodiments. However, any simple modifications, equivalent changes, and alterations made to the above embodiments based on the technical essence of the present invention, without departing from the scope of the present invention, shall still fall within the protection scope of the present invention.

Claims

1. A method for analyzing power meteorological forecasting errors based on transmission line attribute weighting, characterized in that, Includes the following steps: S1: Obtain geographic information data and voltage level attribute data of transmission lines within the target area; obtain meteorological grid prediction data and real-time data within the same historical period and geographic area; S2: For each meteorological grid point, calculate the spatial distance weight, voltage level weight, and tower density weight of the meteorological grid point relative to each adjacent transmission line. Calculate the temporary comprehensive weight coefficient of the meteorological grid point, and take the maximum value as the final comprehensive weight coefficient of the meteorological grid point. The final comprehensive weight coefficients of all meteorological grid points constitute a weight coefficient matrix. The temporary comprehensive weight coefficient of the meteorological grid point is the product of the spatial distance weight, voltage level weight, and tower density weight of the meteorological grid point relative to each adjacent transmission line. Wherein, the spatial distance weight The calculation method is as follows: Calculate the shortest spatial distance from the meteorological grid point to all transmission lines. Spatial distance weights are calculated using a Gaussian function. ; ； Where L is the characteristic attenuation distance. This is the spatial distance weighting correction coefficient; Find the meteorological grid Find the nearest transmission line and obtain its voltage level V. Then, retrieve its voltage level weight by querying a predefined voltage level-weight mapping table. ; Adjust the voltage level weights according to the voltage level weight correction factor: ； in This is a voltage level weighting correction factor; In order to be consistent with meteorological grids The voltage of the nearest transmission line; This is a voltage level-weight mapping function. It is a natural constant; Based on current meteorological grid Centered on a circular statistical area with a radius of K kilometers, the number of all transmission line towers within this area is counted. The tower density weight is calculated using the following formula. : ； in, This is the tower density weighting correction factor; S3: Calculate the error between the predicted data and the actual data at each meteorological grid point; S4: The error value of each meteorological grid point is weighted using the weighting coefficient matrix to calculate the weighted regional error index that reflects the accuracy of meteorological forecasts around the transmission line.

2. The power meteorological forecasting error analysis method based on transmission line attribute weighting according to claim 1, characterized in that, Step S3 specifically involves calculating the values ​​for each meteorological grid point. Mean absolute error Introducing error analysis weighting factors The mean absolute error is corrected, and calculations are performed for each meteorological grid point. The overall prediction error .

3. The power meteorological forecasting error analysis method based on transmission line attribute weighting according to claim 2, characterized in that, The error analysis weighting factors are obtained as follows: Each meteorological grid point is on the [number]th The predicted and actual event values ​​at each time step are converted into binary data based on the event threshold, where 1 indicates that the event has occurred and 0 indicates that the event has not occurred. Thus, for the entire analysis area and time period, a binary field of event predictions and actual event values ​​is obtained: ； Where TS is the event forecast score, H is the number of meteorological grid points that were predicted to occur and actually occurred, FA is the number of meteorological grid points that were predicted to occur but did not actually occur, and M is the number of meteorological grid points that were missed but did not occur. The calculated event forecast score (TS score) is converted into error analysis weighting factors. The conversion rules are as follows: ； in, It is a positive number, used to prevent the denominator from being zero.

4. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the power meteorological forecasting error analysis method based on transmission line attribute weighting as described in any one of claims 1 to 3.

5. An electronic device comprising a memory, a processor, and a computer program stored in the memory, characterized in that, When the processor executes the computer program, it implements the power meteorological forecasting error analysis method based on transmission line attribute weighting as described in any one of claims 1 to 3.

Images