Transformer area load prediction method and device, medium and product

Abstract

The invention discloses a transformer area load prediction method and device, a medium and a product, and relates to the technical field of load prediction, and the method comprises the steps: obtaining single-day historical load curves of a plurality of continuous historical dates before a to-be-predicted day, and carrying out the sample expansion, and obtaining an expanded historical load curve set; determining a historical load per unit value curve and a load base value corresponding to each two-day historical load curve; clustering all the historical load per unit value curves to obtain a plurality of clustering sets; determining all historical load per unit value curves in the same cluster set as the historical load per unit value curve of the historical date before the to-be-predicted day as a second curve; determining a daily load per unit value curve to be predicted based on all the second curves; determining a load base value of the to-be-predicted day based on all the second curves; and determining a predicted load curve based on the daily load per unit value curve to be predicted and the load base value. According to the invention, the accuracy of transformer area short-term load prediction is improved.

Description

Technical Field

[0001] The present application relates to the technical field of load forecasting, and in particular to a method, device, medium and product for predicting load in a substation. Background Art

[0002] Although China's low-voltage power distribution system is large in scale, it is still in its early stages in terms of intelligence, which is mainly reflected in the lack of refined management, especially the accuracy of short-term load forecasting in the substation. However, with the continuous advancement of distribution Internet of Things technology, a large number of distributed terminals have been deployed in the low-voltage distribution network, providing a new way for data analysis of the low-voltage distribution network and opening up an era of distributed measurement with the substation as the basic unit. The terminal can realize autonomous monitoring, management and two-way interaction of the low-voltage distribution network by collecting and analyzing massive operating data in the substation in real time, significantly improving the lean management quality of the low-voltage distribution network.

[0003] As the core link of lean management of power system, load forecasting has evolved from traditional mathematical statistical analysis to machine learning based on neural network, and then to multivariate integrated forecasting. Early mathematical statistical forecasting methods, such as autoregressive model, autoregressive moving average model, differential autoregressive moving average model and volumetric Kalman filter, are widely used because of their simple implementation and small demand for training samples. With the rapid growth of power system load, mathematical statistical algorithms are difficult to accurately capture the nonlinear characteristics of load, resulting in a decrease in prediction accuracy. Entering the 21st century, with the rise of artificial intelligence, load forecasting technology based on neural networks has gradually become a research hotspot. For example, extreme learning machines, artificial neural networks, recurrent neural networks and their integrated applications, the model can automatically learn the inherent laws and patterns in the data, fully consider the complex nonlinear relationship in the power load affected by weather, holidays, temperature and other factors, and significantly improve the prediction accuracy compared with traditional statistical methods. However, when faced with limited training samples (the amount of load data stored in distributed measurement is limited), there may be a problem of large prediction errors. In recent years, in order to solve the above limitations, researchers have begun to explore new prediction methods that integrate multiple data processing technologies to improve the adaptability to system load characteristics. Practice shows that when the number of available samples is small, the use of an integrated forecasting strategy can usually achieve lower forecast errors than a single model. In the integrated forecasting model, how to select historically similar days and determine the data decomposition algorithm become key factors affecting forecast accuracy.

[0004] However, the above technologies mainly focus on the main station-based short-term load forecasting of the substation based on machine learning training big data samples. However, the following problems still exist for the current complex substation short-term load forecasting based on distributed measurement:

[0005] The load at the substation level is small in magnitude, and compared with the system-level load, it has the characteristics of strong randomness and large volatility, which brings certain challenges to its load forecasting work.

[0006] Although distributed measurement equipment has modular functions of measurement, calculation, storage and uploading, it is limited by the amount of data that can be stored in the equipment, which makes the original data used for load forecasting constitute a small sample problem, further increasing the difficulty of load forecasting and making the accuracy of short-term load forecasting in the substation low. Summary of the invention

[0007] The purpose of this application is to provide a method, device, medium and product for predicting load in a substation, thereby improving the accuracy of short-term load prediction in the substation.

[0008] To achieve the above objectives, this application provides the following solutions:

[0009] In a first aspect, the present application provides a method for predicting load in a substation, comprising:

[0010] Obtaining the single-day historical load curves of the substation for multiple consecutive historical dates before the day to be predicted; one historical date corresponds to one single-day historical load curve, the horizontal coordinate of the single-day historical load curve is the historical moment, and the vertical coordinate is the load; the historical moments in different single-day historical load curves are the same and correspond one to one;

[0011] Based on each single-day historical load curve, sample expansion is performed to obtain an expanded historical load curve set; the expanded historical load curve set includes: a plurality of two-day historical load curves; the two-day historical load curve is a curve composed of pre-processed historical load curves of two adjacent historical dates; the number of two-day historical load curves is equal to the number of single-day historical load curves minus 1;

[0012] Based on each of the two-day historical load curves, determine the historical load per unit value curve and the load base value corresponding to each of the two-day historical load curves; the horizontal coordinate of the historical load per unit value curve is the historical time, and the vertical coordinate is the load per unit value;

[0013] Using an affine propagation clustering algorithm, clustering all historical load per-unit value curves to obtain a plurality of cluster sets; the cluster sets include a plurality of historical load per-unit value curves;

[0014] The historical load per unit value curve corresponding to the historical date one day before the forecast date is determined as the first curve, and all the historical load per unit value curves in the same cluster set as the first curve are determined as the second curve;

[0015] Based on all the second curves, a load per unit value curve for the day to be predicted is determined; the load per unit value curve for the day to be predicted has a horizontal coordinate of each time of the day to be predicted, and a vertical coordinate of the load per unit value;

[0016] Determine the load base value of the day to be predicted based on all the second curves and two historical load per unit value curves corresponding to the historical dates three days before the day to be predicted;

[0017] Based on the load per unit value curve of the day to be predicted and the load base value of the day to be predicted, the predicted load curve of the day to be predicted is determined; the horizontal axis of the predicted load curve of the day to be predicted is the various moments of the day to be predicted, and the vertical axis is the load prediction value.

[0018] Optionally, sample expansion is performed based on each single-day historical load curve to obtain an expanded historical load curve set, including:

[0019] Preprocessing each of the single-day historical load curves to obtain a corresponding preprocessed single-day historical load curve;

[0020] The preprocessed single-day historical load curves of any two adjacent dates are spliced ​​to obtain the expanded historical load curve set.

[0021] Optionally, based on each of the two-day historical load curves, determining a historical load per unit value curve and a load base value corresponding to each of the two-day historical load curves includes:

[0022] Determine any two-day historical load curve as the current curve;

[0023] Based on the load at each historical moment in the current curve, determine the load per unit value at each historical moment in the current curve, and obtain a historical load per unit value curve corresponding to the current curve;

[0024] Based on the load and the load per unit value at any historical moment in the current curve, a load base value of the current curve is determined.

[0025] Optionally, based on the load at each historical moment in the current curve, determining the load per unit value at each historical moment in the current curve, and obtaining a historical load per unit value curve corresponding to the current curve, includes:

[0026] Determine any historical moment in the current curve as a target historical moment;

[0027] Based on the load at the target historical moment and the maximum load value and the minimum load value in the current curve, the load per unit value at the target historical moment is determined, thereby obtaining a historical load per unit value curve corresponding to the current curve.

[0028] Optionally, based on all the second curves, determining a per unit value curve of the daily load to be predicted includes:

[0029] Determine the historical load per unit value curves in the second half of each of the second curves as similar daily load per unit value curves;

[0030] Based on all similar daily load per unit value curves, determine the daily load per unit value curve to be predicted.

[0031] Optionally, based on all the second curves and two historical load per unit value curves corresponding to historical dates three days before the day to be predicted, the load base value of the day to be predicted is determined, including:

[0032] The load base value of the day to be predicted is determined based on the load base values ​​of all second curves and the load base values ​​of two historical load per unit value curves corresponding to historical dates three days before the day to be predicted.

[0033] Optionally, determining a forecast load curve for the day to be forecasted based on the load per unit value curve for the day to be forecasted and the load base value for the day to be forecasted includes:

[0034] Multiply the load per unit value at each moment in the load per unit value curve of the day to be predicted by the load base value of the day to be predicted, and obtain the predicted load at each moment of the day to be predicted;

[0035] Based on the predicted load at each time of the day to be predicted, a predicted load curve for the day to be predicted is determined.

[0036] In a second aspect, the present application provides a computer device, comprising: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement any of the above-mentioned methods for predicting load in substations.

[0037] In a third aspect, the present application provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements any of the above-mentioned methods for predicting load in a substation.

[0038] In a fourth aspect, the present application provides a computer program product, including a computer program, which, when executed by a processor, implements any of the above-mentioned methods for predicting load in a substation.

[0039] According to the specific embodiments provided in this application, this application discloses the following technical effects:

[0040] The present application discloses a method, device, medium and product for load prediction of a substation. First, obtain the single-day historical load curves of the substation for multiple consecutive historical dates before the day to be predicted; perform sample expansion based on each single-day historical load curve to obtain an expanded historical load curve set; secondly, determine the historical load per-unit value curve and load base value corresponding to each two-day historical load curve based on each two-day historical load curve; then, cluster all historical load per-unit value curves using an affine propagation clustering algorithm to obtain multiple cluster sets; determine the historical load per-unit value curve corresponding to the historical date one day before the day to be predicted as the first curve, and determine all historical load per-unit value curves in the same cluster set as the first curve as the second curve; then, determine the load per-unit value curve of the day to be predicted based on all the second curves; thirdly, determine the load base value of the day to be predicted based on all the second curves and the two historical load per-unit value curves corresponding to the historical date three days before the day to be predicted; finally, determine the predicted load curve of the day to be predicted based on the load per-unit value curve of the day to be predicted and the load base value of the day to be predicted. The present application decomposes the load into load per unit value and load base value for separate prediction, which can overcome the uncertainty problem of small load magnitude and large fluctuation in the substation, and improve the accuracy of short-term load prediction in the substation. BRIEF DESCRIPTION OF THE DRAWINGS

[0041] In order to more clearly illustrate the technical solutions in the embodiments of the present application or related technologies, the drawings required for use in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0042] Figure 1 A schematic diagram of a flow chart of a method for predicting load in a substation area provided in an embodiment of the present application;

[0043] Figure 2 It is a schematic diagram of the process of realizing the load forecasting method for the substation area;

[0044] Figure 3 It is the per unit value curve of similar daily load;

[0045] Figure 4 It is the per unit value curve of the daily load to be tested;

[0046] Figure 5 It is a comparison chart between the predicted load curve and the actual load curve of the day to be tested;

[0047] Figure 6 A schematic diagram of the structure of a computer device provided in one embodiment of the present application. DETAILED DESCRIPTION

[0048] The following will be combined with the drawings in the embodiments of the present application to clearly and completely describe the technical solutions in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.

[0049] The purpose of this application is to provide a method, device, medium and product for load forecasting in a substation, aiming to improve the accuracy of short-term load forecasting in the substation.

[0050] In order to make the above-mentioned objects, features and advantages of the present application more obvious and easy to understand, the present application is further described in detail below in conjunction with the accompanying drawings and specific implementation methods.

[0051] In an exemplary embodiment, Figure 1 and Figure 2 As shown, the load prediction method of the substation in this embodiment includes:

[0052] Step 1: Obtain the daily historical load curves of the substation for multiple consecutive historical dates before the day to be predicted.

[0053] Among them, one historical date corresponds to a single-day historical load curve, the horizontal axis of the single-day historical load curve is the historical moment, and the vertical axis is the load; the historical moments in different single-day historical load curves are the same and correspond one to one.

[0054] Step 2: Based on each single-day historical load curve, sample expansion is performed to obtain an expanded historical load curve set; the expanded historical load curve set includes: multiple two-day historical load curves.

[0055] Among them, the two-day historical load curve is a curve composed of the pre-processed historical load curves of two adjacent historical dates; the number of two-day historical load curves is equal to the number of single-day historical load curves minus 1.

[0056] As an optional implementation, step 2 includes:

[0057] Step 21: pre-process each single-day historical load curve to obtain the corresponding pre-processed single-day historical load curve.

[0058] Specifically, data preprocessing is performed on the load at each historical moment in each single-day historical load curve. During data preprocessing, the regression method is used to remove bad data points through formulas (1)-(4). Taking a single-day historical load curve X as an example, step 21 is specifically as follows:

[0059] The regression function of the regression method is defined as:

[0060]

[0061]

[0062] Where m(X) is the regression function of the single-day historical load curve X; n is the number of historical moments in the historical load data curve; W(x i ) is the load weight at the i-th historical moment in the historical load data curve; x i is the load at the i-th historical moment in the historical load data curve, i = 1, 2, ..., n; S is the linear smoothing operator of the function m(X); W1(x1) is the self-weight of x1 for the first historical moment; W2(x1) is the mutual weight of x1 for the second historical moment; W n W1(x1) is the mutual weight of x1 to the nth historical moment; W1(x2) is the mutual weight of x2 to the first historical moment; W2(x2) is the self-weight of x2 to the second historical moment; W n (x2) is the mutual weight of x2 to the nth historical moment; W1(x n ) is x n The mutual weight for the first historical moment; W2(x n ) is x n The mutual weight for the second historical moment; W n (x n ) is x n The self-weight for the nth historical moment.

[0063] Assume that the relationship between the single-day historical load curve X and the regression function m(X) is:

[0064] X=m(X)+ε (3)

[0065] Among them, ε is the error term between the single-day load data curve and the regression function, and ε is infinitely close to 0. The uncertainty range of load data prediction, that is, the confidence interval, can be determined by increasing and decreasing the corresponding error range of the regression prediction value. Thus, the upper and lower limits of the prediction value, that is, the confidence interval, are defined as:

[0066]

[0067] Among them, σ is the standard deviation of the data, α=0.1 is taken when calculating the confidence interval, and Z α / 2 =1.645, and the loads at historical moments outside the confidence interval are judged as bad point data and removed.

[0068] Step 22: splice the preprocessed single-day historical load curves of any two adjacent dates to obtain an expanded historical load curve set.

[0069] Specifically, in order to cope with the limitation of the amount of data stored in the distributed measurement equipment itself, the original data used for load forecasting, that is, the single-day historical load curve, constitutes a small sample problem. The joint probability method can be used to process the single-day historical load curve. If the continuous load data of multiple days is regarded as an overall pattern, it can be assumed that the load behavior of the previous period will shape the development trend of the next period of load at the probability level. That is, the load change of each day does not exist in isolation, but has a certain statistical correlation with the load pattern of the previous day or even earlier. By analyzing its correlation, the future load condition can be predicted. Therefore, the pre-processed single-day historical load curves of any two adjacent dates are spliced ​​to obtain the expanded historical load curve set.

[0070] When the historical load curves from the 1st to the 30th day are used to predict the predicted load curve on the 31st day, the expanded historical load curve set includes 29 two-day historical load curves.

[0071] Step 3: Based on each two-day historical load curve, determine the historical load per unit value curve and load base value corresponding to each two-day historical load curve.

[0072] Among them, the horizontal axis of the historical load per unit value curve is the historical moment, and the vertical axis is the load per unit value.

[0073] As an optional implementation, step 3 includes:

[0074] Step 31: Determine any two-day historical load curve as the current curve.

[0075] Step 32: Based on the load at each historical moment in the current curve, determine the load per unit value at each historical moment in the current curve, and obtain the historical load per unit value curve corresponding to the current curve.

[0076] As an optional implementation, step 32 includes:

[0077] Step 321: Determine any historical moment in the current curve as the target historical moment.

[0078] Step 322: Based on the load at the target historical moment and the maximum load value and the minimum load value in the current curve, determine the load per unit value at the target historical moment, thereby obtaining a historical load per unit value curve corresponding to the current curve.

[0079] Specifically, the calculation formula for the load per unit value at any historical moment in any historical load per unit value curve is:

[0080]

[0081] in, is the load per unit value at the i-th historical moment in the historical load per unit value curve; xmax is the maximum load value in the historical load per unit value curve; x min It is the minimum load value in the historical load per unit value curve.

[0082] Step 33: Based on the load and load per unit value at any historical moment in the current curve, determine the load base value of the current curve.

[0083] Specifically, the load at any historical moment in the current curve is divided by the load per unit value of the current curve to obtain the load base value of the current curve.

[0084] Step 4: Use the affine propagation clustering algorithm to cluster all historical load per-unit value curves to obtain multiple cluster sets; the cluster set includes multiple historical load per-unit value curves.

[0085] Step 5: Determine the historical load per unit value curve corresponding to the historical date one day before the day to be predicted as the first curve, and determine all historical load per unit value curves in the same cluster set as the first curve as the second curve.

[0086] Specifically, when the historical load curves from the 1st to the 30th day are used to predict the predicted load curve on the 31st day, the first curve is the 29th historical load per unit value curve.

[0087] Step 6: Based on all the second curves, determine the per unit value curve of the daily load to be predicted.

[0088] Among them, the horizontal axis of the load per unit value curve of the day to be predicted is the various moments of the day to be predicted, and the vertical axis is the load per unit value.

[0089] As an optional implementation, step 6 includes:

[0090] Step 61: Determine the historical load per unit value curves in the second half of each second curve as similar daily load per unit value curves.

[0091] Specifically, the load per unit value curves for similar days are as follows: Figure 3 shown.

[0092] Step 62: Determine the load per unit value curve of the day to be predicted based on all similar daily load per unit value curves.

[0093] Specifically, the per unit value curve of the daily load to be predicted is as follows: Figure 4 As shown in the figure, the calculation formula of the load per unit value at any moment in the load per unit value curve of the predicted day is:

[0094]

[0095] in, is the load per unit value at the i-th moment in the load per unit value curve of the day to be predicted; M is the number of similar daily load per unit value curves; It is the load per unit value at the ith historical moment in the mth similar daily load per unit value curve.

[0096] Step 7: Determine the load base value of the day to be predicted based on all the second curves and the two historical load per unit value curves corresponding to the historical dates three days before the day to be predicted.

[0097] As an optional implementation, step 7 includes:

[0098] The load base value of the day to be predicted is determined based on the load base values ​​of all second curves and the load base values ​​of two historical load per unit value curves corresponding to historical dates three days before the day to be predicted.

[0099] Specifically, the calculation formula for the load base value of the forecast day is:

[0100]

[0101] Among them, Z c is the load base value of the day to be predicted; Z m is the load base value of the mth second curve; Z a is the load base value of the second historical load per unit value curve among the two historical load per unit value curves corresponding to the historical date three days before the forecast date; b It is the load base value of the first historical load per unit value curve of the two historical load per unit value curves corresponding to the historical date three days before the day to be predicted.

[0102] When the historical load curves from the 1st to the 30th day are used to predict the load curve for the 31st day, Z a is the load base value of the 29th historical load per unit value curve, Z b It is the load base value of the 28th historical load per unit value curve.

[0103] Step 8: Based on the load per unit value curve of the day to be predicted and the load base value of the day to be predicted, determine the predicted load curve of the day to be predicted.

[0104] Among them, the horizontal axis of the predicted load curve of the day to be predicted is the various moments of the day to be predicted, and the vertical axis is the load prediction value.

[0105] As an optional implementation, step 8 includes:

[0106] Step 81: Multiply the load per unit value at each moment in the load per unit value curve of the day to be predicted by the load base value of the day to be predicted, to obtain the predicted load at each moment of the day to be predicted.

[0107] Specifically, the calculation formula for the forecast load at any time on the forecast day is:

[0108]

[0109] in, is the forecast load at the i-th moment of the day to be forecasted.

[0110] Step 82: Determine the predicted load curve of the day to be predicted based on the predicted load at each moment of the day to be predicted.

[0111] Specifically, taking the historical load curve from the 1st to the 30th day to predict the predicted load curve on the 31st day as an example, the method of the present application is verified, such as Figure 5 As shown, the predicted load curve obtained by the method of the present application has a high degree of overlap with the actual load curve measured on the 31st day.

[0112] In an exemplary embodiment, a computer device is provided, including: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement a method for predicting loads in a substation.

[0113] In an exemplary embodiment, a computer-readable storage medium is provided, on which a computer program is stored. When the computer program is executed by a processor, a method for predicting load in a substation is implemented.

[0114] In an exemplary embodiment, a computer program product is provided, including a computer program, which implements a method for predicting load in a substation when executed by a processor.

[0115] In an exemplary embodiment, a computer device is provided. The computer device may be a server or a terminal. The internal structure diagram thereof may be as follows: Figure 6 As shown. The computer device includes a processor, a memory, an input / output interface (Input / Output, referred to as I / O) and a communication interface. The processor, the memory and the input / output interface are connected through a system bus, and the communication interface is connected to the system bus through the input / output interface. The processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program and a database. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The input / output interface of the computer device is used to exchange information between the processor and an external device. The communication interface of the computer device is used to communicate with an external terminal through a network connection. When the computer program is executed by the processor, a method for predicting load in a substation is implemented.

[0116] Those skilled in the art will understand that Figure 6 The structure shown in the figure is only a block diagram of a part of the structure related to the solution of the present application, and does not constitute a limitation on the computer device to which the solution of the present application is applied. The specific computer device may include more or fewer components than those shown in the figure, or combine certain components, or have a different arrangement of components.

[0117] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, stored data, displayed data, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties, and the collection, use and processing of relevant data must comply with relevant regulations.

[0118] Those of ordinary skill in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be completed by instructing the relevant hardware through a computer program, and the computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to the memory, database or other medium used in the embodiments provided in the present application can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetic random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. By way of illustration and not limitation, RAM may be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM).

[0119] The database involved in each embodiment provided in this application may include at least one of a relational database and a non-relational database. The non-relational database may include a distributed database based on blockchain, etc., but is not limited thereto. The processor involved in each embodiment provided in this application may be a general-purpose processor, a central processing unit, a graphics processor, a digital signal processor, a programmable logic device, a data processing logic device based on quantum computing, etc., but is not limited thereto.

[0120] The technical features of the above embodiments may be combined arbitrarily. To make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0121] This article uses specific examples to illustrate the principles and implementation methods of this application. The description of the above embodiments is only used to help understand the method and core ideas of this application. At the same time, for those skilled in the art, according to the ideas of this application, there will be changes in the specific implementation methods and application scope. In summary, the content of this specification should not be understood as limiting this application.

Claims

1. A method for predicting load in a substation, characterized in that: The load prediction method for the substation area includes: Obtaining the single-day historical load curves of the substation for multiple consecutive historical dates before the day to be predicted; one historical date corresponds to one single-day historical load curve, the horizontal coordinate of the single-day historical load curve is the historical moment, and the vertical coordinate is the load; the historical moments in different single-day historical load curves are the same and correspond one to one; Based on each single-day historical load curve, sample expansion is performed to obtain an expanded historical load curve set; the expanded historical load curve set includes: a plurality of two-day historical load curves; the two-day historical load curve is a curve composed of pre-processed historical load curves of two adjacent historical dates; the number of two-day historical load curves is equal to the number of single-day historical load curves minus 1; Based on each of the two-day historical load curves, determine the historical load per unit value curve and the load base value corresponding to each of the two-day historical load curves; the horizontal coordinate of the historical load per unit value curve is the historical time, and the vertical coordinate is the load per unit value; Using an affine propagation clustering algorithm, clustering all historical load per-unit value curves to obtain a plurality of cluster sets; the cluster sets include a plurality of historical load per-unit value curves; The historical load per unit value curve corresponding to the historical date one day before the forecast date is determined as the first curve, and all the historical load per unit value curves in the same cluster set as the first curve are determined as the second curve; Based on all the second curves, a load per unit value curve for the day to be predicted is determined; the load per unit value curve for the day to be predicted has a horizontal coordinate of each time of the day to be predicted, and a vertical coordinate of the load per unit value; Determine the load base value of the day to be predicted based on all the second curves and two historical load per unit value curves corresponding to the historical dates three days before the day to be predicted; Based on the load per unit value curve of the day to be predicted and the load base value of the day to be predicted, the predicted load curve of the day to be predicted is determined; the horizontal axis of the predicted load curve of the day to be predicted is the various moments of the day to be predicted, and the vertical axis is the load prediction value.

2. The method for predicting load in a substation according to claim 1, characterized in that: Based on the daily historical load curves, the sample is expanded to obtain the expanded historical load curve set, including: Preprocessing each of the single-day historical load curves to obtain a corresponding preprocessed single-day historical load curve; The preprocessed single-day historical load curves of any two adjacent dates are spliced ​​to obtain the expanded historical load curve set.

3. The method for predicting load in a substation according to claim 1, characterized in that: Based on the two-day historical load curves, respectively, determining the historical load per unit value curve and the load base value corresponding to each of the two-day historical load curves, including: Determine any two-day historical load curve as the current curve; Based on the load at each historical moment in the current curve, determine the load per unit value at each historical moment in the current curve, and obtain a historical load per unit value curve corresponding to the current curve; Based on the load and the load per unit value at any historical moment in the current curve, a load base value of the current curve is determined.

4. The method for predicting load in a substation according to claim 3, characterized in that: Based on the load at each historical moment in the current curve, the load per unit value at each historical moment in the current curve is determined to obtain a historical load per unit value curve corresponding to the current curve, including: Determine any historical moment in the current curve as a target historical moment; Based on the load at the target historical moment and the maximum load value and the minimum load value in the current curve, the load per unit value at the target historical moment is determined, thereby obtaining a historical load per unit value curve corresponding to the current curve.

5. The method for predicting load in a substation according to claim 1, characterized in that: Based on all the second curves, determine the per unit value curve of the daily load to be predicted, including: Determine the historical load per unit value curves in the second half of each of the second curves as similar daily load per unit value curves; Based on all similar daily load per unit value curves, determine the daily load per unit value curve to be predicted.

6. The method for predicting load in a substation according to claim 1, characterized in that: Based on all the second curves and the two historical load per unit value curves corresponding to the historical dates three days before the day to be predicted, the load base value of the day to be predicted is determined, including: The load base value of the day to be predicted is determined based on the load base values ​​of all second curves and the load base values ​​of two historical load per unit value curves corresponding to historical dates three days before the day to be predicted.

7. The method for predicting load in a substation according to claim 1, characterized in that: Based on the load per unit value curve of the day to be predicted and the load base value of the day to be predicted, the predicted load curve of the day to be predicted is determined, including: Multiply the load per unit value at each moment in the load per unit value curve of the day to be predicted by the load base value of the day to be predicted, and obtain the predicted load at each moment of the day to be predicted; Based on the predicted load at each time of the day to be predicted, a predicted load curve for the day to be predicted is determined.

8. A computer device comprising: A memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the substation load forecasting method according to any one of claims 1 to 7.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the method for predicting load in a substation as described in any one of claims 1 to 7 is implemented.

10. A computer program product, comprising a computer program, characterized in that When the computer program is executed by a processor, the method for predicting load in a substation as described in any one of claims 1 to 7 is implemented.

Images