Power grid load prediction method and device and storage medium
By analyzing the meteorological parameters of the power grid area and building a collection of multiple time windows, the problem of poor universality of the existing power grid load prediction methods in different meteorological environments is solved, and more accurate and general grid load prediction is achieved.
Patent Information
- Application Number
- CN202411972461.5
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-12-30
- Publication Date
- 2025-05-30
AI Technical Summary
The existing grid load prediction methods are poor in different meteorological environments. If the model is not retrained, the prediction results will be poor, and different models need to be trained for different meteorological conditions, which is costly.
By determining the meteorological parameters of the area where the power grid is located, multiple time window sets are constructed, the difference values of meteorological parameters are analyzed, the target time window set is determined, and the future maximum load value is predicted based on the historical maximum load value.
It improves the accuracy of grid load prediction, avoids the need to retrain models in different meteorological environments, ensures the universality of prediction methods and saves costs.
Smart Images
Figure CN120067630A_ABST
Abstract
Description
Technical Field
[0001] This application relates to the technical field of data processing, and particularly to a method, device, and storage medium for predicting power grid load. Background Art
[0002] Accurate prediction of power grid load is beneficial to advance power generation and operation mode adjustment, and is of great significance for power supply guarantee during summer peak load and winter peak load periods.
[0003] Currently, the prediction method for power grid load usually trains a prediction model through machine learning based on historical power grid load and meteorological data, and uses the prediction model to predict future power grid load. However, this prediction method usually requires retraining different models for different meteorological environments, with poor generality. If the model is not retrained in different meteorological environments, the prediction result will be poor. Summary of the Invention
[0004] To solve the above technical problems, this application provides a method, device, and storage medium for predicting power grid load, which can improve the accuracy of predicting power grid load.
[0005] In a first aspect, this application provides a method for predicting power grid load, including: determining meteorological parameters corresponding to each time window in a preset time period for the area where the power grid is located; the preset time period includes a future time window and multiple historical time windows; constructing multiple first time window sets and multiple second time window sets according to the time sequence relationship; the first time window set includes the future time window and j historical time windows, and the second time window set includes k + 1 historical time windows; j is greater than a first threshold and less than a second threshold, and k is greater than a third threshold and less than a fourth threshold; based on the meteorological parameters, analyzing the difference values between each first time window set and each second time window set, and determining a target first time window set and a target second time window set according to the difference values; predicting the maximum load value corresponding to the future time window based on the maximum load values corresponding to each historical time window in the target first time window set and the target second time window set.
[0006] In some embodiments, determining the meteorological parameters corresponding to each time window in the first time period for the area where the power grid is located includes: determining the average value of all meteorological parameters within the historical time window as the meteorological parameter corresponding to the historical time window; determining the average value of the meteorological parameters of the future time window predicted by at least one prediction channel as the meteorological parameter corresponding to the future time window.
[0007] In some embodiments, constructing a plurality of first time window sets and a plurality of second time window sets according to a temporal relationship includes: according to the temporal relationship of each time window, combining j consecutive historical time windows whose last time window in the time sequence is a future time window with the future time window to form a first time window set; determining k + 1 consecutive historical time windows in the time sequence as a second time window set.
[0008] In some embodiments, analyzing the difference values between each first time window set and the second time window set includes: forming a first vector from the meteorological parameters corresponding to the j historical time windows and one future time window in the first time window set; forming a second vector from the meteorological parameters corresponding to the k + 1 historical time windows in the second time window set; constructing a difference value matrix template according to the dimensions of the first vector and the second vector; recursively calculating the elements in the difference value matrix template based on the first vector, the second vector, and the Euclidean distance between the first vector and the second vector to obtain a difference value matrix; and determining the value in the upper right corner of the difference value matrix as the difference value between the first time window set and the second time window set.
[0009] In some embodiments, recursively calculating the elements in the difference value matrix template based on the first vector, the second vector, and the Euclidean distance between the first vector and the second vector to obtain a difference value matrix includes: recursively calculating the first element in the first row and first column of the difference value matrix template according to the formula recursively calculating each element in the first column according to the formula recursively calculating each element in the (j + 1)-th row according to the formula recursively calculating the elements in other positions according to the formula where x s is used to represent the eigenvalue in the first vector, y t is used to represent the eigenvalue in the second vector; m is used to represent the number of historical time windows in a preset time period; is used to represent the first element in the first row and first column, is used to represent the element in the s-th row and first column, is used to represent the element in the (j + 1)-th row and t-th column, is used to represent the element in the s-th row and t-th column; |x - y| is used to represent the Euclidean distance between the first vector and the second vector.
[0010] In some embodiments, predicting the maximum load value corresponding to a future time window based on the maximum load values corresponding to each historical time window in the target first time window set and the target second time window set includes: determining, from the target first time window set, n historical time windows whose time sequences are consecutive with the future time window as the first historical time window set, and determining a first sum value; the first sum value is the sum of the maximum load values corresponding to each historical time window in the first historical time window set; determining, from the target second time window set, the historical time window with the last time sequence as the most similar historical time window, and determining the most similar load value; the most similar load value is the maximum load value corresponding to the most similar historical time window; determining, from the target second time window set, n historical time windows whose time sequences are consecutive with the most similar historical time window as the second historical time window set, and determining a second sum value; the second sum value is the sum of the maximum load values corresponding to each historical time window in the second historical time window set; determining the ratio between the first sum value and the second sum value, and determining the product of the ratio and the maximum load value as the maximum load value corresponding to the future time window.
[0011] In some embodiments, j takes at least one value from 2, 3, 4, 5, 6, 7, 8; k takes at least one value from 2, 3, 4, 5, 6, 7, 8.
[0012] In a second aspect, the present application provides a prediction device for grid load, including: a determination module, configured to determine the meteorological parameters corresponding to each time window within a preset time period in the area where the power grid is located; the preset time period includes a future time window and multiple historical time windows; a construction module, configured to construct multiple first time window sets and multiple second time window sets according to the time sequence relationship; the first time window set includes the future time window and j historical time windows, and the second time window set includes k + 1 historical time windows; j is greater than a first threshold and less than a second threshold, and k is greater than a third threshold and less than a fourth threshold; an analysis module, configured to analyze the difference value between each first time window set and each second time window set based on the meteorological parameters, and determine a target first time window set and a target second time window set according to the difference value; a prediction module, configured to predict the maximum load value corresponding to the future time window based on the maximum load values corresponding to each historical time window in the target first time window set and the target second time window set.
[0013] In some embodiments, the determination module is specifically configured to: determine the average value of all meteorological parameters within the historical time window as the meteorological parameter corresponding to the historical time window; determine the average value of the meteorological parameters of the future time window predicted by at least one prediction channel as the meteorological parameter corresponding to the future time window.
[0014] In some embodiments, the construction module is specifically configured to: according to the timing relationship of each time window, form a first time window set by combining j consecutive historical time windows whose last time window in terms of timing is a future time window and the future time window; determine k + 1 consecutive historical time windows in terms of timing as the second time window set.
[0015] In some embodiments, the analysis module includes a combination sub-module, a construction sub-module, a recurrence sub-module, and a determination sub-module; the combination sub-module is configured to form a first vector by combining the meteorological parameters corresponding to the j historical time windows in the first time window set and a future time window; form a second vector by combining the meteorological parameters corresponding to the k + 1 historical time windows in the second time window set; the construction sub-module is configured to construct a difference value matrix template according to the dimensions of the first vector and the second vector; the recurrence sub-module is configured to recur the elements in the difference value matrix template according to the first vector, the second vector, and the Euclidean distance between the first vector and the second vector to obtain a difference value matrix; the determination sub-module is configured to determine the value in the upper right corner of the difference value matrix as the difference value between the first time window set and the second time window set.
[0016] In some embodiments, the recurrence sub-module is specifically configured to: according to the formula recursively obtain the first element in the first row and the first column in the difference value matrix template; according to the formula recursively obtain each element in the first column; according to the formula recursively obtain each element in the (j + 1)-th row; according to the formula recursively obtain the elements in other positions in the difference value matrix template; where x s is used to represent the eigenvalue in the first vector, y t is used to represent the eigenvalue in the second vector; m is used to represent the number of historical time windows in a preset time period; is used to represent the first element in the first row and the first column, is used to represent the element in the s-th row and the first column, is used to represent the element in the (j + 1)-th row and the t-th column, is used to represent the element in the s-th row and the t-th column; |x - y| is used to represent the Euclidean distance between the first vector and the second vector.
[0017] In some embodiments, the prediction module is specifically configured to: determine n historical time windows with consecutive time sequences in the target first time window set and the future time window as the first historical time window set, and determine the first sum value; the first sum value is the sum of the maximum load values corresponding to each historical time window in the first historical time window set; determine the historical time window with the last time sequence in the target second time window set as the most similar historical time window, and determine the most similar load value; the most similar load value is the maximum load value corresponding to the most similar historical time window; determine n historical time windows with consecutive time sequences to the most similar historical time window in the target second time window set as the second historical time window set, and determine the second sum value; the second sum value is the sum of the maximum load values corresponding to each historical time window in the second historical time window set; determine the ratio between the first sum value and the second sum value, and determine the product of the ratio and the maximum load value as the maximum load value corresponding to the future time window.
[0018] In some embodiments, j takes at least one value from 2, 3, 4, 5, 6, 7, 8; k takes at least one value from 2, 3, 4, 5, 6, 7, 8.
[0019] In a third aspect, the present application provides an electronic device, including: a processor, a memory, and a computer program stored on the memory and executable on the processor, where when the computer program is executed by the processor, it implements the power grid load prediction method in any embodiment of the first aspect.
[0020] In a fourth aspect, the present application provides a computer-readable storage medium, including: a computer program stored on the computer-readable storage medium, where when the computer program is executed by the processor, it implements the power grid load prediction method in any embodiment of the first aspect.
[0021] In a fifth aspect, the present application provides a computer program product, including: when the computer program product runs on a computer, it causes the computer to implement the power grid load prediction method in any embodiment of the first aspect.
[0022] The technical solution provided by this application has the following advantages compared with the prior art: First, determine the meteorological parameters corresponding to each time window in the preset time period for the area where the power grid is located. Among them, the preset time period includes a future time window and multiple historical time windows. Then, construct multiple first time window sets and multiple second time window sets according to the chronological relationship. Among them, the first time window set includes the future time window and j historical time windows, and the second time window set includes k + 1 historical time windows; j is greater than the first threshold and less than the second threshold, and k is greater than the third threshold and less than the fourth threshold. Then, based on the meteorological parameters, analyze the difference values between each first time window set and the second time window set, and determine the target first time window set and the target second time window set according to the difference values. Finally, based on the maximum load values corresponding to each historical time window in the target first time window set and the target second time window set, predict the maximum load value corresponding to the future time window. In this way, it is possible to first determine the historical time period set with the smallest difference from the time period set including the future time period according to the meteorological parameters of the area where the power grid is located, and then calculate the maximum load value of the future time period according to the maximum load values corresponding to the time periods in the historical time period set and the time periods in the time period set including the future time period. It avoids the problem that when predicting the power grid load by using the training model, if the model is not retrained in different meteorological environments, the prediction result will be poor, and improves the accuracy of the prediction result. At the same time, there is no need to train different models for different meteorological conditions, which ensures the generality of the prediction method and saves costs at the same time. BRIEF DESCRIPTION OF THE DRAWINGS
[0023] The accompanying drawings herein are incorporated into the specification and constitute a part of this specification, showing embodiments consistent with the present application and, together with the specification, are used to explain the principles of the present application.
[0024] To more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the following will briefly introduce the accompanying drawings required for the description of the embodiments or the prior art. Obviously, for those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative efforts.
[0025] Figure 1 It is a schematic diagram of an application scenario of a method for predicting the power grid load provided by an embodiment of the present application;
[0026] Figure 2 It is one of the flow diagrams of the method for predicting the power grid load provided by an embodiment of the present application;
[0027] Figure 3 It is another flow diagram of the method for predicting the power grid load provided by an embodiment of the present application;
[0028] Figure 4 It is the third flow schematic diagram of the power grid load prediction method provided by the embodiment of the present application;
[0029] Figure 5 It is the structural schematic diagram of a power grid load prediction device provided by the embodiment of the present application;
[0030] Figure 6 It is the structural schematic diagram of an electronic device provided by the embodiment of the present application. Detailed implementation manners
[0031] In order to be able to more clearly understand the above-mentioned objects, features and advantages of the present application, the solution of the present application will be further described below. It should be noted that, without conflict, the embodiments of the present application and the features in the embodiments can be combined with each other.
[0032] Many specific details are set forth in the following description in order to fully understand the present application, but the present application may also be implemented in other ways different from those described herein; obviously, the embodiments in the specification are only a part of the embodiments of the present application, rather than all of the embodiments.
[0033] A stable power supply is a necessary condition for social production and life. In recent years, with global climate change, extreme high temperature and low temperature events have occurred frequently, posing a serious threat to the reliable supply of power grid load. Accurate prediction of the daily maximum load of the power grid is conducive to the power grid to carry out power generation and operation mode adjustment in advance, and is of great significance for power supply guarantee during the peak summer and winter periods. At present, the prediction method for power grid load usually uses historical power grid load and meteorological data to train a prediction model through machine learning, and uses the prediction model to predict the future power grid load. However, this prediction method usually requires retraining different models for different meteorological environments, and the generality is poor. If the model is not retrained in different meteorological environments, the prediction result will be expected to be poor.
[0034] In view of the above problems, the embodiments of the present application provide a method for predicting the power grid load. First, determine the meteorological parameters corresponding to each time window in a preset time period for the area where the power grid is located, and construct a plurality of first time window sets and a plurality of second time window sets according to the time sequence relationship. Then, based on the meteorological parameters, analyze the difference values between each first time window set and the second time window sets, and determine the target first time window set and the target second time window set according to the difference values. Finally, based on the maximum load values corresponding to each historical time window in the target first time window set and the target second time window set, predict the maximum load value corresponding to the future time window. In this way, it is possible to first determine the historical time period set with the smallest difference from the time period set including the future time period according to the meteorological parameters of the area where the power grid is located, and then calculate the maximum load value of the future time period according to the maximum load values corresponding to the time periods in the historical time period set and the maximum load values corresponding to the time periods in the time period set including the future time period. This avoids the problem that when predicting the power grid load by using a training model, if the model is not retrained in different meteorological environments, the prediction result will be poor, and improves the accuracy of the prediction result. At the same time, there is no need to train different models for different meteorological conditions, which ensures the generality of the prediction method and saves costs at the same time.
[0035] The power grid load prediction method provided by the embodiments of the present disclosure is mainly applicable to the scenario of predicting the maximum load of a power grid affected by meteorological conditions. Figure 1 This is an application scenario diagram of a power grid load prediction method provided by the embodiments of the present disclosure. As Figure 1 shown, the application scenario includes a terminal 102 and a server 104. Among them, the terminal 102 can communicate with the server 104 through a network. The terminal 102 can include, but is not limited to, various personal computers, laptop computers, smart phones, tablet computers, etc., and the server 104 can be implemented by an independent server or a server cluster composed of multiple servers.
[0036] It should be noted that the power grid load prediction method provided by the embodiments of the present application can be executed by an electronic device. The electronic device can be the server 104, or the terminal 102, or a device integrating the functions of the server 104 and the terminal 102. The present application does not make any limitations in this regard.
[0037] The power grid load prediction method provided by the embodiments of this application can be executed by a power grid load prediction device, which can be hardware or software. When the power grid load prediction device is hardware, it can be various electronic devices with the function of predicting the power grid load, including but not limited to mobile phones, computers, laptops, tablets, etc. When the power grid load prediction device is software, it can be installed in the above-listed electronic devices. It can be implemented as multiple software or software modules, or as a single software or software module. No specific limitation is made here.
[0038] Figure 2 It is a schematic flowchart of the power grid load prediction method provided by the embodiments of this application. As Figure 2 shown, the power grid load prediction method may include the following steps:
[0039] S11. Determine the meteorological parameters corresponding to each time window in the preset time period for the area where the power grid is located.
[0040] Among them, the preset time period includes a future time window and multiple historical time windows. The length of the historical time window is the same as that of the future time window. For example, the length of the historical time window and the future time window can be one day, one week, one month, etc. The length of the preset time period is an integer multiple of the time window length. For example, when the time window length is 1 day, the length of the preset time period can be 365 days. In some embodiments, the length of the preset time period is at least greater than or equal to one year.
[0041] Specifically, the method for determining the meteorological parameters corresponding to each time window in the preset time period for the area where the power grid is located can be to determine the average value of all meteorological parameters in the historical time window as the meteorological parameters corresponding to the historical time window; and determine the average value of the meteorological parameters of the future time window predicted by at least one prediction channel as the meteorological parameters corresponding to the future time window. The prediction channel is any channel capable of predicting future time meteorological parameters, such as different websites, different applications (APPs), etc.
[0042] In some embodiments, the meteorological parameters at least include temperature, humidity, wind speed, and precipitation.
[0043] S12. Construct multiple first time window sets and multiple second time window sets according to the time sequence relationship.
[0044] Among them, the first time window set includes a future time window and j historical time windows, and the second time window set includes k + 1 historical time windows. j is greater than the first threshold and less than the second threshold; k is greater than the third threshold and less than the fourth threshold; both j and k are integers. The first threshold, the second threshold, the third threshold, and the fourth threshold are all preset. For example, they are default values, or values set by relevant personnel according to the actual situation. In some embodiments, j takes at least one of 2, 3, 4, 5, 6, 7, 8; k takes at least one of 2, 3, 4, 5, 6, 7, 8.
[0045] Specifically, the manner of constructing multiple first time window sets and multiple second time window sets according to the chronological relationship can be to form a first time window set by combining the j consecutive historical time windows whose last time window in the chronological order is the future time window and the future time window according to the chronological relationship of each time window; and determine k + 1 consecutive historical time windows in the chronological order as the second time window set.
[0046] Exemplarily, when the time windows within a preset time period are sorted chronologically as A, B, C, D, E, F, G, and G is the future time window, j = 3, and k = 5, the first time window set includes E, F, G; the second time window set includes A, B, C, D, E, or B, C, D, E, F.
[0047] S13. Analyze the difference values between each first time window set and each second time window set based on meteorological parameters, and determine a target first time window set and a target second time window set according to the difference values.
[0048] First, analyze the difference values between each first time window set and each second time window set based on meteorological parameters.
[0049] Specifically, the manner of analyzing the difference values between each first time window set and each second time window set based on meteorological parameters can be to first horizontally calculate the first average value of the meteorological parameters in the first time window set, horizontally calculate the second average value of the meteorological parameters in the second time window set, and then vertically determine the absolute value of the difference between the first average value and the second average value as the difference value between the first time window set and the second time window set. Among them, when there are multiple meteorological parameters, when calculating the first average value, the sum of the average values of the multiple meteorological parameters in the first time window set is determined as the first average value. Similarly, when there are multiple meteorological parameters, when calculating the second average value, the sum of the average values of the multiple meteorological parameters in the second time window set is determined as the second average value.
[0050] The method for analyzing the difference values between each first time window set and the second time window set based on meteorological parameters can also be to first form two different vectors from the meteorological parameters corresponding to the first time window set and the second time window set, and then construct a difference matrix based on the dimensions and Euclidean distances of the two different vectors, and determine the value in the upper right corner of the difference matrix (i.e., the first value in the last column of the first row) as the difference value between the first time window set and the second time window set.
[0051] Secondly, determine the target first time window set and the target second time window set according to the difference values. Specifically, determine the first time window set corresponding to the smallest difference value as the target first time window set, and determine the second time window set with the smallest difference value as the target second time window set.
[0052] S14. Predict the maximum load value corresponding to the future time window based on the maximum load values corresponding to each historical time window in the target first time window set and the target second time window set.
[0053] Specifically, the method for predicting the maximum load value corresponding to the future time window based on the maximum load values corresponding to each historical time window in the target first time window set and the target second time window set can be to directly determine the maximum load value corresponding to the historical time window with the last sequence in the target second time window set as the maximum load value corresponding to the future time window. It can also be obtained by weighting the maximum load value corresponding to the historical time window with the last sequence in the target second time window set according to the maximum load values corresponding to n historical time windows in the target first time window set and the maximum load values corresponding to n historical time windows in the target second time window set. Among them, the n historical time windows in the target second time window set do not include the historical time window with the last sequence, n ≤ j, and n ≤ k.
[0054] In the above solution, first, determine the meteorological parameters corresponding to each time window in the preset time period for the area where the power grid is located. The preset time period includes a future time window and multiple historical time windows. Then, construct multiple first time window sets and multiple second time window sets according to the chronological relationship. The first time window set includes the future time window and j historical time windows, and the second time window set includes k + 1 historical time windows; j is greater than the first threshold and less than the second threshold, and k is greater than the third threshold and less than the fourth threshold. Then, based on the meteorological parameters, analyze the difference values between each first time window set and the second time window set, and determine the target first time window set and the target second time window set according to the difference values. Finally, based on the maximum load values corresponding to each historical time window in the target first time window set and the target second time window set, predict the maximum load value corresponding to the future time window. In this way, it is possible to first determine, according to the meteorological parameters of the area where the power grid is located, the historical time period set with the smallest difference from the time period set including the future time period, and then calculate the maximum load value of the future time period according to the maximum load values corresponding to the time periods in the historical time period set and the maximum load values corresponding to the time periods in the time period set including the future time period. This avoids the problem that when predicting the power grid load by using a training model, if the model is not retrained in different meteorological environments, the prediction result will be poor, and improves the accuracy of the prediction result. At the same time, there is no need to train different models for different meteorological conditions, which ensures the generality of the prediction method and saves costs at the same time.
[0055] In some embodiments, as Figure 3 shown, the method for analyzing the difference values between each first time window set and the second time window set based on the meteorological parameters may include the following steps:
[0056] S131. Combine the meteorological parameters of the j historical time windows and one future time window in the first time window set to form a first vector.
[0057] Exemplarily, denote the meteorological parameters (temperature, precipitation, wind speed, humidity) corresponding to the future time window as T 0 , P 0 , W 0 , H 0 , and let a 0 = [T 0 , P 0 , W 0 , H 0 ; Denote the meteorological parameters corresponding to the first historical time window chronologically adjacent to the future time window as T 1 , P 1 , W 1 , H 1, let a 1 = [T 1 , P 1 , W 1 , H 1 , and so on. The meteorological parameters corresponding to the historical time window separated from the future time window by 1 + i time window lengths are respectively denoted as T 1+i , P 1+i , W 1+i , H 1+i , let a 1+i = [T 1+i , P 1+i , W 1+i , H 1+i , where i = 1, 2, 3,..., N; then the first vector can be expressed as: X j = [x 1 , x 2 ,..., x j , x j+1 = [a j , a j-1 ,..., a 1 , a 0 , where j = 2, 3, 4, 5, 6, 7, 8.
[0058] S132. Combine the meteorological parameters corresponding to k + 1 historical time windows in the second time window set to form a second vector.
[0059] Similar to step S131, for example, denote the meteorological parameters (temperature, precipitation, wind speed, humidity) corresponding to the future time window as T 0 , P 0 , W 0 , H 0 , and let a 0 = [T 0 , P 0 , W 0 , H 0 ; denote the meteorological parameters corresponding to the first historical time window adjacent to the future time window in time sequence as T 1 , P 1 , W 1 , H 1 , and let a 1 = [T 1 , P 1 , W 1 , H 1 , and so on. The meteorological parameters corresponding to the historical time window separated from the future time window by 1 + i time window lengths are respectively denoted as T 1+i , P 1+i , W 1+i , H 1+i, let a 1+i = [T 1+i , P 1+i , W 1+i , H 1+i , where i = 1, 2, 3,..., N; then the second vector can be expressed as: where k = 3, 4, 5, 6, 7; m = 1, 2, 3,..., N - k.
[0060] S133. Construct a difference value matrix template according to the dimensions of the first vector and the second vector.
[0061] Specifically, the first vector is a (j + 1)-dimensional vector, and the second vector is a (k + 1)-dimensional vector. Therefore, a difference value matrix template with dimensions of (j + 1)×(k + 1) is constructed according to the dimensions of the first vector and the second vector.
[0062] S134. Recursively calculate the elements in the difference value matrix template based on the first vector, the second vector, and the Euclidean distance between the first vector and the second vector to obtain the difference value matrix.
[0063] Specifically, according to the first vector X j and the second vector Let |x - y| represent the Euclidean distance between vectors x and y, and recursively calculate the elements in the difference value matrix template based on the first vector, the second vector, and the Euclidean distance between the first vector and the second vector to obtain the difference value matrix.
[0064] The elements in the difference value matrix template can be recursively calculated through the following A - D:
[0065] A. According to the formula Recursively obtain the first element in the first row and the first column of the difference value matrix template; s = 2, 3, 4,..., j + 1.
[0066] B. According to the formula Recursively obtain each element in the first column of the difference value matrix template; t = 2, 3, 4,..., k + 1.
[0067] C. According to the formula Recursively obtain each element in the (j + 1)-th row of the difference value matrix template; s = 2, 3, 4,..., j + 1; t = 2, 3, 4,..., k + 1.
[0068] D. According to the formula Recursively obtain the elements in other positions of the difference value matrix template.
[0069] where x s is used to represent the eigenvalue in the first vector, y tused to characterize the eigenvalues in the second vector; m is used to characterize the number of historical time windows in the preset time period; used to characterize the first element in the first row and the first column, used to characterize the element in the s-th row and the first column, used to characterize the element in the (j + 1)-th row and the t-th column, used to characterize the element in the s-th row and the t-th column; |x - y| is used to characterize the Euclidean distance between the first vector and the second vector.
[0070] In the above solution, it is possible to dynamically deduce the elements in the entire difference value matrix by combining the Euclidean distance between the first vector and the second vector, solving the problem that it is difficult to evaluate the difference between vectors of different lengths, and indirectly improving the accuracy of the prediction result. At the same time, it is ensured that regardless of how the meteorological environment changes, the difference between every two vectors of different lengths can be dynamically and flexibly deduced, improving the generality of the charge prediction method.
[0071] S135. Determine the value in the upper right corner of the difference value matrix as the difference value between the first time window set and the second time window set.
[0072] Specifically, when the determined difference value matrix is the difference value matrix in step S134, let Determine the value in the upper right corner of the difference value matrix as the difference value between the first time window set and the second time window set. Among them, used to characterize the difference value, used to characterize the value of the first element in the last column of the first row of the difference value matrix (i.e., the value in the upper right corner of the difference value matrix).
[0073] In the above solution, it is possible to construct a multi-dimensional difference value matrix template based on two vectors of different lengths, and deduce and fill the elements in each position of the template according to the Euclidean distance between the vectors, and finally obtain the difference value between the two vectors, that is, the difference value between the first time window set and the second time window set. In this way, on the one hand, a method for evaluating the difference between vectors of different lengths is provided, ensuring the accuracy of the prediction result. On the other hand, it is also ensured that regardless of how the meteorological environment changes, the difference between every two vectors of different lengths can be dynamically and flexibly deduced, improving the generality of the charge prediction method.
[0074] In some embodiments, as Figure 4 shown, based on the maximum load values corresponding to each historical time window in the target first time window set and the target second time window set, the method for predicting the maximum load value corresponding to the future time window may include the following steps:
[0075] S141. Determine the first historical time window set as the n historical time windows with continuous time series and future time windows in the target first time window set, and determine the first sum value.
[0076] Among them, the first sum value is the sum of the maximum load values corresponding to each historical time window in the first historical time window set, n ≤ j, and n ≤ k.
[0077] S142. Determine the historical time window with the last time series in the target second time window set as the most similar historical time window, and determine the most similar load value.
[0078] Among them, the most similar load value is the maximum load value corresponding to the most similar historical time window.
[0079] S143. Determine the second historical time window set as the n historical time windows with continuous time series and the most similar historical time window in the target second time window set, and determine the second sum value.
[0080] Among them, the second sum value is the sum of the maximum load values corresponding to each historical time window in the second historical time window set.
[0081] S144. Determine the ratio between the first sum value and the second sum value, and determine the product of the ratio and the maximum load value as the maximum load value corresponding to the future time window.
[0082] Specifically, the maximum load value corresponding to the future time window can be calculated according to the following formula.
[0083]
[0084] Among them, L 0 is used to represent the maximum load value corresponding to the future time window; L q is used to represent the most similar load value, L 1 is used to represent the sum of the maximum load values corresponding to the n historical time windows with continuous time series and the future time window in the target first time window set; L 2 is used to represent the sum of the maximum load values corresponding to the n historical time windows with continuous time series and the most similar historical time window in the target second time window set.
[0085] In the above solution, the maximum load value of the future time period can be calculated according to the maximum load value corresponding to the time period in the historical time period set and the maximum load value corresponding to the time period in the time period set including the future time period. It avoids the problem that when predicting the power grid load by using the training model, if the model is not retrained in different meteorological environments, the prediction result will be poor, and improves the accuracy of the prediction result.
[0086] The embodiments of the present application can divide the functional modules of the power grid load prediction device according to the above method examples. For example, each functional module can be divided corresponding to each function, or two or more functions can be integrated into one processing unit. The above integrated module can be implemented in the form of hardware or in the form of a software functional module. It should be noted that the division of modules in the embodiments of the present application is illustrative, only a logical function division, and there can be other division methods in actual implementation.
[0087] As Figure 5 shown, it is a schematic structural diagram of the power grid load prediction device provided by the embodiment of the present application. The power grid load prediction device includes a determination module 51, a construction module 52, an analysis module 53, and a prediction module 54.
[0088] The determination module 51 is used to determine the meteorological parameters corresponding to each time window in a preset time period in the area where the power grid is located; the preset time period includes a future time window and multiple historical time windows; the construction module 52 is used to construct multiple first time window sets and multiple second time window sets according to the time sequence relationship; the first time window set includes the future time window and j historical time windows, and the second time window set includes k + 1 historical time windows; j is greater than a first threshold and less than a second threshold, and k is greater than a third threshold and less than a fourth threshold; the analysis module 53 is used to analyze the difference values between each first time window set and the second time window set based on the meteorological parameters, and determine the target first time window set and the target second time window set according to the difference values; the prediction module 54 is used to predict the maximum load value corresponding to the future time window based on the maximum load values corresponding to each historical time window in the target first time window set and the target second time window set.
[0089] In the above solution, first, determine the meteorological parameters corresponding to each time window in the preset time period for the area where the power grid is located. The preset time period includes a future time window and multiple historical time windows. Then, construct multiple first time window sets and multiple second time window sets according to the chronological relationship. The first time window set includes the future time window and j historical time windows, and the second time window set includes k + 1 historical time windows; j is greater than the first threshold and less than the second threshold, and k is greater than the third threshold and less than the fourth threshold. Then, based on the meteorological parameters, analyze the difference values between each first time window set and the second time window set, and determine the target first time window set and the target second time window set according to the difference values. Finally, based on the maximum load values corresponding to each historical time window in the target first time window set and the target second time window set, predict the maximum load value corresponding to the future time window. In this way, it is possible to first determine the set of historical time periods with the smallest difference from the set of time periods including the future time period according to the meteorological parameters of the area where the power grid is located, and then calculate the maximum load value of the future time period according to the maximum load values corresponding to the time periods in the set of historical time periods and the maximum load values corresponding to the time periods in the set of time periods including the future time period. This avoids the problem that when predicting the power grid load by using a training model, if the model is not retrained in different meteorological environments, the prediction result will be poor, and improves the accuracy of the prediction result. At the same time, there is no need to train different models for different meteorological conditions, which ensures the generality of the prediction method and saves costs at the same time.
[0090] In some embodiments, the determining module 51 is specifically configured to: determine the average value of all meteorological parameters within the historical time window as the meteorological parameter corresponding to the historical time window; determine the average value of the meteorological parameters of the future time window predicted by at least one prediction channel as the meteorological parameter corresponding to the future time window.
[0091] In some embodiments, the constructing module 52 is specifically configured to: according to the chronological relationship of each time window, form a first time window set by combining the consecutive j historical time windows with the future time window where the last time window in the time sequence is the future time window; determine the consecutive k + 1 historical time windows in the time sequence as the second time window set.
[0092] In some embodiments, the analysis module 53 includes a combination sub-module, a construction sub-module, a recurrence sub-module, and a determination sub-module; the combination sub-module is configured to form a first vector by combining the meteorological parameters corresponding to j historical time windows and one future time window in the first time window set; and form a second vector by combining the meteorological parameters corresponding to k + 1 historical time windows in the second time window set; the construction sub-module is configured to construct a difference value matrix template according to the dimensions of the first vector and the second vector; the recurrence sub-module is configured to recur the elements in the difference value matrix template according to the first vector, the second vector, and the Euclidean distance between the first vector and the second vector to obtain a difference value matrix; the determination sub-module is configured to determine the value in the upper right corner of the difference value matrix as the difference value between the first time window set and the second time window set.
[0093] In some embodiments, the recurrence sub-module is specifically configured to: recur to obtain the first element in the first row and the first column in the difference value matrix template according to the formula recur to obtain each element in the first column in the difference value matrix template according to the formula recur to obtain each element in the (j + 1)-th row in the difference value matrix template according to the formula recur to obtain the elements in other positions in the difference value matrix template according to the formula where x s is used to represent the eigenvalue in the first vector, y t is used to represent the eigenvalue in the second vector; m is used to represent the number of historical time windows in the preset time period; is used to represent the first element in the first row and the first column, is used to represent the element in the s-th row and the first column, is used to represent the element in the (j + 1)-th row and the t-th column, is used to represent the element in the s-th row and the t-th column; |x - y| is used to represent the Euclidean distance between the first vector and the second vector.
[0094] In some embodiments, the prediction module 54 is specifically configured to: determine n historical time windows with continuous time series with respect to the future time window in the target first time window set as the first historical time window set, and determine the first sum value; the first sum value is the sum of the maximum load values corresponding to each historical time window in the first historical time window set; determine the historical time window with the last time series in the target second time window set as the most similar historical time window, and determine the most similar load value; the most similar load value is the maximum load value corresponding to the most similar historical time window; determine n historical time windows with continuous time series with respect to the most similar historical time window in the target second time window set as the second historical time window set, and determine the second sum value; the second sum value is the sum of the maximum load values corresponding to each historical time window in the second historical time window set; determine the ratio between the first sum value and the second sum value, and determine the product of the ratio and the maximum load value as the maximum load value corresponding to the future time window.
[0095] In some embodiments, j takes at least one value from 2, 3, 4, 5, 6, 7, 8; k takes at least one value from 2, 3, 4, 5, 6, 7, 8.
[0096] The power grid load prediction device provided in this embodiment can execute the power grid load prediction method provided in the above method embodiment, and its implementation principle and technical effect are similar to those of the above method, and will not be elaborated here.
[0097] Figure 6 is an electronic device shown according to an exemplary embodiment. The electronic device may include a processor 802, and the processor 802 is configured to execute application program code to implement the power grid load prediction method in the present application.
[0098] The processor 802 may be a central processing unit (CPU), a microprocessor, an application-specific integrated circuit (ASIC), or one or more integrated circuits for controlling the execution of the program of the solution of the present application.
[0099] As Figure 6 shown, the electronic device may further include a memory 803. Among them, the memory 803 is used to store the application program code for executing the solution of the present application and is controlled by the processor 802 to execute.
[0100] The memory 803 can be a read-only memory (ROM) or other types of static storage devices that can store static information and instructions, a random access memory (RAM) or other types of dynamic storage devices that can store information and instructions, or can also be an electrically erasable programmable read-only memory (EEPROM), a compact disc read-only memory (CD-ROM) or other optical disc storage, optical disc storage (including compact discs, laser discs, optical discs, digital versatile discs, Blu-ray discs, etc.), magnetic disk storage media or other magnetic storage devices, or any other medium that can be used to carry or store the desired program code in the form of instructions or data structures and can be accessed by a computer, but is not limited thereto. The memory 803 can exist independently and be connected to the processor 802 through the bus 804. The memory 803 can also be integrated with the processor 802.
[0101] As Figure 6 shown, the electronic device can also include a communication interface 801. Among them, the communication interface 801, the processor 802, and the memory 803 can be coupled to each other. For example, they can be coupled to each other through the bus 804. The communication interface 801 is used for information interaction with other devices. For example, it supports information interaction between the electronic device and other devices.
[0102] It should be noted that Figure 6 the device structure shown in Figure 6 does not constitute a limitation on the electronic device. In addition to Figure 6 the components shown, the electronic device can include more or fewer components than those shown in the figure, or combine certain components, or have different component arrangements. And the electronic device provided in this embodiment can execute the power grid load prediction method provided in the above method embodiment. The implementation principle and technical effect are similar to those of the above method and will not be elaborated here.
[0103] This application embodiment provides a computer-readable storage medium. A computer program is stored on the computer-readable storage medium. When the computer program is executed by a processor, it realizes each process of the power grid load prediction method in the above method embodiment and can achieve the same technical effect. To avoid repetition, it will not be elaborated here.
[0104] Among them, the computer-readable storage medium can be a read-only memory (ROM), a random access memory (RAM), a magnetic disk, or an optical disc, etc.
[0105] An embodiment of the present application provides a computer program product. The computer program product stores a computer program, and when the computer program is executed by a processor, it implements each process of the power grid load prediction method in the above method embodiment and can achieve the same technical effect. To avoid repetition, it will not be elaborated here.
[0106] Those skilled in the art should understand that the embodiments of the present application can be provided as a method, a system, or a computer program product. Therefore, the present application can adopt the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware aspects. Moreover, the present application can adopt the form of a computer program product implemented on one or more computer-usable storage media that contain computer-usable program code.
[0107] In the present application, the processor can be a central processing unit (CPU), or other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor can be a microprocessor or the processor can also be any conventional processor, etc.
[0108] In the present application, the memory may include non-permanent memory in a computer-readable medium, random access memory (RAM), and / or non-volatile memory in the form of, for example, read-only memory (ROM) or flash RAM. The memory is an example of a computer-readable medium.
[0109] In this application, computer-readable media include both permanent and non-permanent, removable and non-removable storage media. The storage media can implement information storage by any method or technology, and the information can be computer-readable instructions, data structures, program modules, or other data. Examples of computer storage media include, but are not limited to, phase change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technologies, compact disc read-only memory (CD-ROM), digital versatile disc (DVD) or other optical storage, magnetic cassette tapes, disk storage or other magnetic storage devices, or any other non-transmission media that can be used to store information accessible by a computing device. As defined herein, computer-readable media do not include transitory media such as modulated data and carrier waves.
[0110] It should be noted that, in this article, relational terms such as "first" and "second" are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any actual relationship or order between these entities or operations. Moreover, the term "comprising", "including" or any other variant thereof is intended to cover non-exclusive inclusion, so that a process, method, article or device comprising a series of elements not only includes those elements, but also includes other elements not expressly listed, or also includes elements inherent to such process, method, article or device. Without further limitation, an element defined by the statement "comprising one..." does not exclude the existence of additional identical elements in the process, method, article or device comprising the said element.
[0111] The above are only specific embodiments of this application, enabling those skilled in the art to understand or implement this application. Various modifications to these embodiments will be obvious to those skilled in the art, and the general principles defined herein can be implemented in other embodiments without departing from the spirit or scope of this application. Therefore, this application will not be limited to these embodiments described herein, but rather will conform to the broadest scope consistent with the principles and novel features disclosed herein.
Claims
1. A method for predicting power grid load, characterized in that: include: Determine the meteorological parameters corresponding to each time window in a preset time period in the area where the power grid is located; the preset time period includes a future time window and multiple historical time windows; Constructing multiple first time window sets and multiple second time window sets according to the time series relationship; the first time window set includes a future time window and j historical time windows, and the second time window set includes k+1 historical time windows; j is greater than a first threshold and less than a second threshold, and k is greater than a third threshold and less than a fourth threshold; Based on the meteorological parameters, analyzing the difference between each first time window set and the second time window set, and determining a target first time window set and a target second time window set according to the difference; Based on the maximum load value corresponding to each historical time window in the target first time window set and the target second time window set, the maximum load value corresponding to the future time window is predicted.
2. The prediction method according to claim 1, characterized in that: The step of determining the meteorological parameters corresponding to each time window in the first time period in the area where the power grid is located includes: Determine the average value of all meteorological parameters within the historical time window as the meteorological parameter corresponding to the historical time window; An average value of the meteorological parameter of the future time window predicted by at least one prediction channel is determined as the meteorological parameter corresponding to the future time window.
3. The prediction method according to claim 1, characterized in that: The step of constructing a plurality of first time window sets and a plurality of second time window sets according to the time sequence relationship includes: According to the time sequence relationship of each time window, the j consecutive historical time windows whose last time window in the time sequence is the future time window are combined with the future time window to form the first time window set; k+1 historical time windows that are continuous in time sequence are determined as a second time window set.
4. The prediction method according to claim 1, characterized in that: The analyzing the difference between each first time window set and each second time window set based on the meteorological parameter comprises: The meteorological parameters corresponding to j historical time windows and one future time window in the first time window set are combined into a first vector; The meteorological parameters corresponding to the k+1 historical time windows in the second time window set are combined into a second vector; constructing a difference value matrix template according to the dimension of the first vector and the dimension of the second vector; recursively deducing elements in the difference value matrix template according to the first vector, the second vector, and the Euclidean distance between the first vector and the second vector to obtain a difference value matrix; The value at the upper right corner of the difference value matrix is determined as the difference value between the first time window set and the second time window set.
5. The prediction method according to claim 4, characterized in that: According to the first vector, the second vector, and the Euclidean distance between the first vector and the second vector, recursively derive elements in the difference value matrix template to obtain a difference value matrix including: According to the formula Recursively obtain the first element of the first row and first column in the difference value matrix template; According to the formula Recursively obtain each element of the first column in the difference value matrix template; According to the formula Recursively obtain each element in the j+1th row of the difference value matrix template; According to the formula Recursively obtain elements at other positions in the difference value matrix template; Among them, x s It is used to characterize the eigenvalues in the first vector, y t is used to characterize the eigenvalues in the second vector; m is used to characterize the number of historical time windows in the preset time period; Used to represent the first element of the first row and first column, Used to represent the element in the first column of the sth row, It is used to represent the element in the j+1th row and tth column. Used to represent the element in row s and column t; | xy | The Euclidean distance between the first vector and the second vector is used to represent the Euclidean distance between the first vector and the second vector.
6. The prediction method according to claim 1, characterized in that: The predicting the maximum load value corresponding to the future time window based on the maximum load value corresponding to each historical time window in the target first time window set and the target second time window set includes: Determine n historical time windows in the target first time window set whose time sequence is continuous with the future time window as a first historical time window set, and determine a first sum value; the first sum value is the sum of the maximum load values corresponding to each historical time window in the first historical time window set; Determine the last historical time window in the target second time window set as the most similar historical time window, and determine the most similar load value; the most similar load value is the maximum load value corresponding to the most similar historical time window; Determine n historical time windows in the target second time window set whose time sequence is continuous with the most similar historical time window as a second historical time window set, and determine a second sum value; the second sum value is the sum of the maximum load values corresponding to each historical time window in the second historical time window set; A ratio between the first sum value and the second sum value is determined, and a product of the ratio and the maximum load value is determined as the maximum load value corresponding to the future time window.
7. The prediction method according to any one of claims 1 to 6, characterized in that: The value of j is at least one of 2, 3, 4, 5, 6, 7, and 8; the value of k is at least one of 2, 3, 4, 5, 6, 7, and 8.
8. A device for predicting power grid load, characterized in that: include: A determination module, used to determine the meteorological parameters corresponding to each time window in a preset time period in the area where the power grid is located; the preset time period includes a future time window and multiple historical time windows; A construction module, configured to construct a plurality of first time window sets and a plurality of second time window sets according to a time series relationship; the first time window set includes a future time window and j historical time windows, and the second time window set includes k+1 historical time windows; j is greater than a first threshold and less than a second threshold, and k is greater than a third threshold and less than a fourth threshold; An analysis module, configured to analyze the difference between each first time window set and the second time window set based on the meteorological parameter, and determine a target first time window set and a target second time window set according to the difference; A prediction module is used to predict the maximum load value corresponding to the future time window based on the maximum load value corresponding to each historical time window in the target first time window set and the target second time window set.
9. An electronic device, characterized in that: include: A processor, a memory, and a computer program stored in the memory and executable on the processor, wherein when the computer program is executed by the processor, the method for predicting power grid load according to any one of claims 1 to 7 is implemented.
10. A computer-readable storage medium, characterized in that: include: The computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the method for predicting power grid load according to any one of claims 1 to 7 is implemented.
11. A computer program product, characterized in that When the computer program product is executed on a computer, the computer is enabled to implement the method for predicting power grid load according to any one of claims 1 to 7.