A load prediction method, device, equipment and computer readable storage medium

Abstract

This invention discloses a load forecasting method, which includes the following steps: decomposing target load power data with mixed noise components using empirical mode decomposition (EMD) to obtain a detrended subsequence of the target quantity; determining the K-value parameter range of variational mode decomposition (VMD) based on the target quantity; determining the optimal decomposition parameters of VMD by combining each detrended subsequence and the K-value parameter range; and predicting the load value corresponding to the target load power data based on the optimal decomposition parameters of VMD. Applying the load forecasting method provided by this invention achieves effective noise reduction of load power data with mixed noise components, significantly improving the accuracy of load forecasting. This invention also discloses a load forecasting device, equipment, and storage medium, which have corresponding technical effects.

Description

Technical Field

[0001] This invention relates to the field of data noise reduction technology, and in particular to a load forecasting method, apparatus, device, and computer-readable storage medium. Background Technology

[0002] Industrial load constitutes the largest proportion of electricity consumption. For industrial enterprises, formulating reasonable power purchase plans based on load forecast curves for a future period can reduce waste, lower electricity costs, achieve demand-side management, and alleviate the supply-demand imbalance in the power system. Noise significantly affects the accuracy of load power data and reduces the forecasting precision of industrial enterprise loads. Therefore, reducing the noise content of industrial enterprise load data is of great significance for improving the accuracy of industrial load forecasting.

[0003] Decomposing and reconstructing load power data with mixed noise components can effectively filter out noise components. Current methods for load power data decomposition mainly include Empirical Mode Decomposition (EMD) and Variational Mode Decomposition (VMD). EMD can decompose non-stationary signals into more stationary component signals, but it is prone to endpoint effects and mode aliasing. Variational Mode Decomposition is a novel decomposition method for non-stationary signals, improving upon the shortcomings of EMD. However, it also suffers from the drawback that the value of the decomposition parameter K cannot be adaptively determined. The selection of K significantly impacts the decomposition performance of Variational Mode Decomposition; theoretically, K cannot be too small or too large. In the former case, a small number of subsequences K is insufficient to fully represent the signal, and some important parts of the signal will be discarded as noise, resulting in severe distortion of the reconstructed signal. In the latter case, an excessively large number of subsequences K usually leads to over-decomposition, which not only fails to effectively filter out noise components but also causes mode aliasing. The inability to effectively reduce noise in load power data with mixed noise components affects the accuracy of load forecasting.

[0004] In summary, how to effectively solve the problem of not being able to effectively reduce noise in load power data with mixed noise components, thus affecting the accuracy of load forecasting, is a problem that urgently needs to be solved by those skilled in the art. Summary of the Invention

[0005] The purpose of this invention is to provide a load forecasting method that effectively reduces noise in load power data with mixed noise components, thereby significantly improving the accuracy of load forecasting. Another purpose of this invention is to provide a load forecasting device, equipment, and computer-readable storage medium.

[0006] To solve the above-mentioned technical problems, the present invention provides the following technical solution:

[0007] A load forecasting method, comprising:

[0008] Empirical mode decomposition is used to decompose the target load power data with mixed noise components to obtain the detrended subsequence of the target quantity;

[0009] The range of K-value parameters for variational mode decomposition is determined based on the target quantity;

[0010] The optimal decomposition parameters for variational mode decomposition are determined by combining the detrended subsequences and the range of K values.

[0011] The load value corresponding to the target load power data is predicted based on the optimal decomposition parameters of the variational mode decomposition.

[0012] In one specific embodiment of the present invention, determining the range of K-value parameters for variational mode decomposition based on the target quantity includes:

[0013] The positive integers between two and the target number are defined as the range of K-value parameters for the variational mode decomposition.

[0014] In one specific embodiment of the present invention, determining the optimal decomposition parameters for variational mode decomposition by combining each of the detrended subsequences and the range of the K-value parameter includes:

[0015] Summation is performed on each of the detrended subsequences to obtain the sum sequence;

[0016] Each positive integer in the range of K values ​​is determined as a target decomposition parameter;

[0017] For each target decomposition parameter, the sum sequence is decomposed into subsequences corresponding to the target decomposition parameter.

[0018] Calculate the decomposition residuals based on the sum sequence and each of the subsequences;

[0019] Calculate the complexity index of the decomposed residual;

[0020] The maximum value among the complexity indices is determined, and the target decomposition parameter corresponding to the maximum value is determined as the optimal decomposition parameter of the variational mode decomposition.

[0021] In one specific embodiment of the present invention, predicting the load value corresponding to the target load power data based on the optimal decomposition parameters of the variational mode decomposition includes:

[0022] Obtain the target subsequences obtained by decomposing the sum sequence using the optimal decomposition parameters of the variational mode decomposition;

[0023] Each of the target subsequences is input into a preset load prediction model to perform load prediction and obtain the load value corresponding to the target load power data.

[0024] In one specific embodiment of the present invention, empirical mode decomposition is used to decompose the target load power data of mixed noise components, including:

[0025] The target load power data of the mixed noise component is sampled at a preset time interval to obtain the power data of each sampled load.

[0026] The empirical mode decomposition is used to decompose each of the sampled load power data.

[0027] In one specific embodiment of the present invention, after sampling the target load power data of the mixed noise components at preset time intervals to obtain the sampled load power data, the method further includes:

[0028] Determine if there is a target time point where the sampling result is empty;

[0029] If so, retrieve the target sampled load power data corresponding to the target time point in the previous sampling period;

[0030] The target sampled load power data is determined as the sampled load power data corresponding to the target time point in this sampling period.

[0031] A load forecasting device, comprising:

[0032] The data decomposition module is used to decompose the target load power data with mixed noise components using empirical mode decomposition to obtain the target number of detrended subsequences.

[0033] The parameter range determination module is used to determine the range of K-value parameters for variational mode decomposition based on the target quantity.

[0034] The optimal decomposition parameter determination module is used to determine the optimal decomposition parameters of variational mode decomposition by combining each of the detrended subsequences and the range of the K-value parameter.

[0035] The load prediction module is used to predict the load value corresponding to the target load power data based on the optimal decomposition parameters of the variational mode decomposition.

[0036] In one specific embodiment of the present invention, the optimal decomposition parameter determination module includes:

[0037] The sequence summation calculation submodule is used to sum the detrended subsequences to obtain the sum sequence;

[0038] The decomposition parameter determination submodule is used to determine each positive integer in the range of K-value parameters as the target decomposition parameter;

[0039] The sequence decomposition submodule is used to decompose the sum sequence into subsequences of the target decomposition parameter for each target decomposition parameter;

[0040] A decomposed residual calculation submodule is used to calculate the decomposed residual based on the sum sequence and each of the subsequences;

[0041] The complexity index calculation submodule is used to calculate the complexity index of the decomposed residual;

[0042] The optimal decomposition parameter determination submodule is used to determine the maximum value among the complexity indicators and determine the target decomposition parameter corresponding to the maximum value as the optimal decomposition parameter of the variational mode decomposition.

[0043] A load forecasting device, comprising:

[0044] Memory, used to store computer programs;

[0045] A processor for executing the computer program to implement the steps of the load forecasting method as described above.

[0046] A computer-readable storage medium storing a computer program that, when executed by a processor, implements the steps of the load forecasting method as described above.

[0047] The load forecasting method provided by this invention uses empirical mode decomposition to decompose the target load power data of mixed noise components to obtain a detrended subsequence of the target quantity; determines the range of K-value parameters for variational mode decomposition based on the target quantity; determines the optimal decomposition parameters for variational mode decomposition by combining each detrended subsequence and the range of K-value parameters; and predicts the load value corresponding to the target load power data based on the optimal decomposition parameters for variational mode decomposition.

[0048] As can be seen from the above technical solution, by first decomposing the target load power data of mixed noise components using empirical mode decomposition (EMD), determining the range of K-value parameters for variational mode decomposition (VMD) based on the target number of the obtained detrended subsequences, and further combining each detrended subsequence with the K-value parameter range to determine the optimal VMD decomposition parameters, the optimal VMD decomposition parameters can be determined. This enables load forecasting based on the optimal VMD decomposition parameters, effectively reducing noise in the load power data of mixed noise components and significantly improving the accuracy of load forecasting.

[0049] Accordingly, the present invention also provides a load forecasting device, equipment, and computer-readable storage medium corresponding to the above-described load forecasting method, which have the above-described technical effects, and will not be elaborated further here. Attached Figure Description

[0050] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0051] Figure 1 This is a flowchart illustrating one implementation of the load forecasting method in this invention.

[0052] Figure 2 This is a flowchart illustrating another implementation of the load forecasting method in this invention.

[0053] Figure 3 This is a line chart illustrating a complexity index in an embodiment of the present invention;

[0054] Figure 4 This is a structural block diagram of a load forecasting device according to an embodiment of the present invention;

[0055] Figure 5 This is a structural block diagram of a load forecasting device according to an embodiment of the present invention;

[0056] Figure 6 This is a schematic diagram of the specific structure of a load forecasting device provided in this embodiment. Detailed Implementation

[0057] To enable those skilled in the art to better understand the present invention, the invention will be further described in detail below with reference to the accompanying drawings and specific embodiments. Obviously, the described embodiments are merely some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0058] See Figure 1 , Figure 1 This is a flowchart illustrating one implementation of the load forecasting method in this invention. The method may include the following steps:

[0059] S101: Empirical mode decomposition is used to decompose the target load power data of mixed noise components to obtain the detrended subsequence of the target quantity.

[0060] When load forecasting is required based on target load power data with mixed noise components, noise reduction processing of the target load power data with mixed noise components is necessary first. Empirical Mode Decomposition (EMD) is used to decompose the target load power data with mixed noise components into a detrended subsequence of the target quantity. In other words, EMD can adaptively decompose the target load power data with mixed noise components into a detrended subsequence of the target quantity.

[0061] S102: Determine the range of K-value parameters for variational mode decomposition based on the target quantity.

[0062] After decomposing the target load power data of the mixed noise components to obtain the target number of detrended subsequences, the range of the K-value parameter for variational mode decomposition is determined based on the target number. The K-value parameter for variational mode decomposition is generally not less than 2 and not greater than the total number of detrended subsequences obtained from decomposing the target load power data of the mixed noise components.

[0063] S103: Determine the optimal decomposition parameters for variational mode decomposition by combining each detrended subsequence and the range of K-value parameters.

[0064] After determining the range of K-value parameters for variational mode decomposition, the optimal decomposition parameters for variational mode decomposition are determined by combining each detrending subsequence with the range of K-value parameters.

[0065] Different values ​​of the K-value parameter in Variational Mode Decomposition (VM) result in varying denoising effects on target load power data with mixed noise components. The K-value parameter in VM determines the number of subsequences obtained by further decomposing the summation of each detrended subsequence. As the number of decompositions increases, the complexity index initially increases and then decreases. Based on the denoising effect of each K-value parameter within its range, the K-value parameter with the best denoising effect can be selected as the optimal decomposition parameter for VM.

[0066] S104: Predict the load value corresponding to the target load power data based on the optimal decomposition parameters of variational mode decomposition.

[0067] After determining the optimal decomposition parameters of the variational mode decomposition (VMD), the load value corresponding to the target load power data is predicted based on these parameters. A load forecasting model can be pre-set, decomposing the summation result of each detrended subsequence into subsequences with the optimal VMD parameters. These subsequences are then input into the load forecasting model, which is used to perform load forecasting to obtain the load value corresponding to the target load power data. This invention combines empirical mode decomposition (EMD) algorithms with improvements to the VMD algorithm, overcoming the limitation of VMD's inability to adaptively optimize decomposition. It achieves adaptive noise reduction for industrial enterprise load forecasting, ensuring excellent VMD decomposition performance while filtering out noise components from the load power data.

[0068] As can be seen from the above technical solution, by first decomposing the target load power data of mixed noise components using empirical mode decomposition (EMD), determining the range of K-value parameters for variational mode decomposition (VMD) based on the target number of the obtained detrended subsequences, and further combining each detrended subsequence with the K-value parameter range to determine the optimal VMD decomposition parameters, the optimal VMD decomposition parameters can be determined. This enables load forecasting based on the optimal VMD decomposition parameters, effectively reducing noise in the load power data of mixed noise components and significantly improving the accuracy of load forecasting.

[0069] It should be noted that, based on the above embodiments, the present invention also provides corresponding improvements. In subsequent embodiments, steps identical or corresponding to those in the above embodiments can be referenced interchangeably, and their respective beneficial effects can also be referred to each other. These improvements will not be elaborated upon in the following improved embodiments.

[0070] See Figure 2 , Figure 2 This is a flowchart illustrating another implementation of the load forecasting method in this invention. The method may include the following steps:

[0071] S201: Sample the target load power data of the mixed noise components according to a preset time interval to obtain the sampled load power data.

[0072] When load forecasting is required based on target load power data of mixed noise components, the time interval for sampling the target load power data of mixed noise components can be preset. The target load power data of mixed noise components is sampled according to the preset time interval to obtain the sampled load power data.

[0073] The target load power data can be represented by γ(t), specifically as follows:

[0074] γ(t)={<γ1,t1>,<γ2,t2>,…,<γ j ,t j >, ..., <γ n ,t n >};

[0075] Where, <γ j ,t j > indicates the load power value and sampling time of the mixed noise component corresponding to sampling point j, and n indicates the number of sampling points. The sampling method is sampling at equal time intervals.

[0076] It should be noted that the sampling time interval for the target load power data can be set and adjusted according to the actual situation, and the embodiments of the present invention do not limit this.

[0077] S202: Determine if there is a target time point with an empty sampling result. If yes, proceed to step S203; otherwise, proceed to step S205.

[0078] During the recording and transmission of signal data, a small amount of data loss is inevitable. After sampling the target load power data of the mixed noise component according to the preset time interval, and obtaining the sampled load power data, it is determined whether there is a target time point with an empty sampling result. If so, it means that no data was collected at the target time point or the sampling result was lost during the transmission process, and step S203 is executed. If not, it means that when sampling according to the preset time interval, the corresponding data can be collected normally at each time point, and step S205 is executed.

[0079] S203: Retrieve the target sampled load power data corresponding to the target time point in the previous sampling period.

[0080] The sampling period for load power data is preset. When it is determined that there is a target time point with no sampling result, it means that no data was collected at the target time point or the collection result was lost during transmission. The target sampled load power data corresponding to the target time point in the previous sampling period is retrieved.

[0081] S204: Determine the target sampled load power data as the sampled load power data corresponding to the target time point in this sampling period.

[0082] After retrieving the target sampled load power data corresponding to the target time point in the previous sampling period, the target sampled load power data is determined as the sampled load power data corresponding to the target time point in the current sampling period. This allows for filling in the missing time values ​​in the collected load power data with signal data from the same time point in the previous week, ensuring data integrity.

[0083] S205: Empirical mode decomposition is used to decompose the power data of each sampled load to obtain the target number of detrended subsequences.

[0084] The target quantity K is adaptively obtained by decomposing γ(t) using empirical mode decomposition. E A trend subsequence In addition, a trend sequence γ can be obtained. q (t).

[0085] S206: Determine the range of positive integers between the two and the target quantity as the range of K-value parameters for variational mode decomposition.

[0086] After decomposing the sampled load power data using empirical mode decomposition to obtain the detrended subsequence of the target number, the positive integers between 2 and the target number, i.e., [2, K], are then used. E The range of K-value parameters for variational mode decomposition is determined. This allows all positive integers between the two and the target quantity to be traversed, enabling the selection of the optimal solution from the positive integers corresponding to the K-values.

[0087] S207: Summate each detrended subsequence to obtain the sum sequence.

[0088] After decomposing the sampled load power data using Empirical Mode Decomposition (EMD) to obtain the target number of detrended subsequences, the summation of each detrended subsequence is performed to obtain the total sum sequence. Define γ′(t) as K E The sum of detrended subsequences can be specifically represented as:

[0089]

[0090] This completes the summation calculation for each detrended subsequence.

[0091] S208: Determine each positive integer in the range of K values ​​as the target decomposition parameter.

[0092] After determining the positive integers between two and the target quantity as the range of K-value parameters for variational mode decomposition, and summing each detrended subsequence to obtain the sum sequence, each positive integer in the range of K-value parameters is determined as the target decomposition parameter.

[0093] S209: For each target decomposition parameter, decompose the sum sequence into target decomposition parameter subsequences.

[0094] By determining each positive integer in the range of K values ​​as the target decomposition parameter, the sum sequence is decomposed into target decomposition parameter subsequences for each target decomposition parameter.

[0095] The steps of decomposing a sum sequence into a target number of subsequences may include:

[0096] (1) Construct a constrained variational problem, described by the following equation:

[0097]

[0098] Where t represents time, u k Let ω represent the k-th subsequence. k Indicate u k The center frequency, δ(t) represents the derivative with respect to t, K is the total number of subsequences, δ(t) is the Ulker function, * is the convolution operator, and st is an abbreviation for subject to, meaning "to make... satisfy...".

[0099] (2) Construct the Lagrange multiplier equation and solve the equation:

[0100]

[0101] Where λ is the Lagrange multiplier factor, used to maintain the strictness of the constraints, and α is the bandwidth control parameter, used to ensure the reconstruction accuracy of the signal.

[0102] (3) Obtained using the alternating direction multiplier method The updated value is expressed by the following equation:

[0103]

[0104] Where N is the number of iterations, and They represent γ′(t), and λ N The Fourier transform of (t), where ω is the angular frequency of the signal.

[0105] (4) Determine whether the subsequence after the update iteration satisfies the conditions of the following convergence equation. If it does not satisfy the conditions, continue the update iteration. If it does satisfy the conditions, end the iteration and obtain the subsequence after sequence decomposition using variational mode decomposition.

[0106]

[0107] Where ε is the convergence criterion tolerance value.

[0108] S210: Calculate the decomposition residuals based on the sum sequence and each subsequence.

[0109] After decomposing the sum sequence into subsequences corresponding to each target decomposition parameter, the decomposition residuals are calculated based on the sum sequence and each subsequence. The decomposition residual S is calculated using the following formula.K (t):

[0110]

[0111] Among them, u k (t) is the subsequence obtained by variational mode decomposition of γ′(t), and K is the number of subsequences.

[0112] Decomposition of residual S K (t) is specifically represented as follows:

[0113] S K (t)={<s1,t1>,<s2,t2>,…,<s j ,t j >,…,<s n ,t n >};

[0114] Among them, <s j ,t j > indicates the decomposition residual value and sampling time corresponding to sampling point j, n indicates the number of sampling points, and the sampling method is sampling at equal time intervals.

[0115] S211: The complexity index for calculating the decomposition residual.

[0116] After calculating the decomposition residuals based on the sum sequence and each subsequence, the complexity index of the decomposition residuals is calculated.

[0117] The specific steps for calculating the complexity index include:

[0118] (1) For the decomposition residuals of variational mode decomposition, construct the m-dimensional vector space S of the decomposition residuals. m (t), specifically represented as follows:

[0119]

[0120] in,

[0121]

[0122] (2) Calculate vectors sum vector Chebyshev distance

[0123]

[0124] (3) Define the similarity threshold function θ(x,M), which is expressed as follows:

[0125]

[0126] In the formula, M is the similarity threshold.

[0127] (4) Calculate vectors with vector similarity Specifically, it is expressed as follows:

[0128]

[0129] (5) Calculate the vector space S m The logarithm B of the average similarity between pairwise vectors within (t) m :

[0130]

[0131] (6) Construct the m+1 dimensional vector space S of the decomposed residuals. m+1 (t), repeat steps (1) to (5) to obtain the vector space S. m+1 The logarithm B of the average similarity between pairwise vectors within (t) m+1 :

[0132]

[0133] (7) Calculate the decomposition residual S K The complexity index X of (t) K :

[0134] X K =B m -B m+1 ;

[0135] Complexity index X K Size is positively correlated with noise level.

[0136] S212: Determine the maximum value among the various complexity indices, and determine the target decomposition parameter corresponding to the maximum value as the optimal decomposition parameter for variational mode decomposition.

[0137] After calculating the complexity index for each objective decomposition parameter, the maximum value among the complexity indices is determined, and the objective decomposition parameter corresponding to the maximum value is determined as the optimal decomposition parameter for variational mode decomposition. This ensures that sequence decomposition is performed using the decomposition parameters K of variational mode decomposition.

[0138] See Figure 3 , Figure 3 This is a line chart illustrating a complexity index in an embodiment of the present invention. (From...) Figure 3Analysis shows that as the number of decompositions increases, the complexity index first increases and then decreases. When the complexity index reaches its maximum value, K=16, it means that the decomposition residual at this time contains all the noise of the detrended sequence, and the detrended sequence can be decomposed most effectively. The optimal number of decomposition layers for variational mode decomposition is 16.

[0139] S213: Obtain the target subsequences obtained by decomposing the sum sequence with the optimal decomposition parameters through variational mode decomposition.

[0140] After determining the target decomposition parameter corresponding to the maximum value as the optimal decomposition parameter of variational mode decomposition, the target subsequence with the optimal decomposition parameter obtained by variational mode decomposition of the sum sequence is obtained.

[0141] S214: Input each target subsequence into the preset load prediction model to perform load prediction using the preset load prediction model and obtain the load value corresponding to the target load power data.

[0142] After obtaining the target subsequences by decomposing the sum sequence with the optimal decomposition parameters through variational mode decomposition, each target subsequence is input into a preset load prediction model to perform load prediction and obtain the load value corresponding to the target load power data.

[0143] This invention introduces a complexity index to measure the denoising effect of the load power data time-domain reconstruction process by using the adaptive optimal decomposition stage of variational mode decomposition, thus performing adaptive denoising processing on the load power data. This invention overcomes the shortcoming of variational mode decomposition algorithms in their inability to adaptively perform optimal decomposition and denoising of load power data. It not only ensures the excellent signal decomposition performance of variational mode decomposition but also filters out noise components in the load power data. This invention can be applied to load power data denoising scenarios in industrial enterprises. This enables load forecasting with optimal decomposition parameters, significantly improving the accuracy of load forecasting.

[0144] This embodiment differs from Embodiment 1 corresponding to the technical solution claimed in Independent Claim 1, and also adds the technical solutions claimed in Dependent Claims 2 to 6. Of course, depending on the actual situation and requirements, the technical solutions claimed in each dependent claim can be flexibly combined without affecting the integrity of the solution, so as to better meet the requirements of different usage scenarios. This embodiment only provides one solution that includes the most solutions and has the best effect. Because the situation is complex, it is impossible to list all possible solutions. Those skilled in the art should realize that there can be many examples based on the basic method principle provided in this application combined with the actual situation. Without sufficient creative effort, they should all be within the protection scope of this application.

[0145] Corresponding to the above method embodiments, the present invention also provides a load forecasting device, and the load forecasting device described below can be referred to in correspondence with the load forecasting method described above.

[0146] See Figure 4 , Figure 4 This is a structural block diagram of a load forecasting device according to an embodiment of the present invention. The device may include:

[0147] Data decomposition module 41 is used to decompose the target load power data with mixed noise components using empirical mode decomposition to obtain the detrended subsequence of the target quantity;

[0148] The parameter range determination module 42 is used to determine the range of K-value parameters for variational mode decomposition based on the number of targets.

[0149] The optimal decomposition parameter determination module 43 is used to determine the optimal decomposition parameters of variational mode decomposition by combining each detrended subsequence and the range of K-value parameters.

[0150] The load prediction module 44 is used to predict the load value corresponding to the target load power data based on the optimal decomposition parameters of variational mode decomposition.

[0151] As can be seen from the above technical solution, by first decomposing the target load power data of mixed noise components using empirical mode decomposition (EMD), determining the range of K-value parameters for variational mode decomposition (VMD) based on the target number of the obtained detrended subsequences, and further combining each detrended subsequence with the K-value parameter range to determine the optimal VMD decomposition parameters, the optimal VMD decomposition parameters can be determined. This enables load forecasting based on the optimal VMD decomposition parameters, effectively reducing noise in the load power data of mixed noise components and significantly improving the accuracy of load forecasting.

[0152] In one specific embodiment of the present invention, the parameter range determination module 42 is specifically a module that determines the positive integer between two and the target quantity as the range of K-value parameters for variational mode decomposition.

[0153] In one specific embodiment of the present invention, the optimal decomposition parameter determination module 43 includes:

[0154] The sequence summation calculation submodule is used to sum the subsequences of each detrended subsequence to obtain the sum sequence;

[0155] The decomposition parameter determination submodule is used to determine each positive integer in the range of K-value parameters as the target decomposition parameter;

[0156] The sequence decomposition submodule is used to decompose the sum sequence into target decomposition parameter number of subsequences for each target decomposition parameter;

[0157] The decomposition residual calculation submodule is used to calculate the decomposition residual based on the sum sequence and each subsequence;

[0158] The complexity metric calculation submodule is used to calculate the complexity metric of the decomposed residuals.

[0159] The optimal decomposition parameter determination submodule is used to determine the maximum value among the various complexity indices and to determine the target decomposition parameter corresponding to the maximum value as the optimal decomposition parameter for variational mode decomposition.

[0160] In one specific embodiment of the present invention, the load forecasting module 44 includes:

[0161] The subsequence acquisition submodule is used to acquire each target subsequence obtained by decomposing the sum sequence through the optimal decomposition parameters of variational mode decomposition.

[0162] The load forecasting submodule is used to input each target subsequence into the preset load forecasting model, so as to use the preset load forecasting model to perform load forecasting and obtain the load value corresponding to the target load power data.

[0163] In one specific embodiment of the present invention, the data decomposition module 41 includes:

[0164] The sampling submodule is used to sample the target load power data of the mixed noise components at preset time intervals to obtain the power data of each sampled load.

[0165] The data decomposition submodule is used to decompose the sampled load power data using empirical mode decomposition.

[0166] In one specific embodiment of the present invention, the device may further include:

[0167] The judgment module is used to sample the target load power data of the mixed noise component at a preset time interval, and after obtaining the sampled load power data, determine whether there is a target time point with an empty sampling result.

[0168] The data retrieval module is used to retrieve the target sampled load power data corresponding to the target time point in the previous sampling period when it is determined that there is a target time point with empty sampling results.

[0169] The data determination module is used to determine the target sampled load power data as the sampled load power data corresponding to the target time point in the current sampling period.

[0170] For the method embodiments described above, see [link to relevant documentation]. Figure 5 , Figure 5 This is a schematic diagram of the load forecasting device provided by the present invention, which may include:

[0171] Memory 332 is used to store computer programs;

[0172] The processor 322 is used to implement the steps of the load forecasting method in the above method embodiments when executing a computer program.

[0173] For details, please refer to Figure 6 , Figure 6 This is a schematic diagram illustrating the specific structure of a load forecasting device provided in this embodiment. The load forecasting device can vary significantly due to differences in configuration or performance. It may include a processor (central processing unit, CPU) 322 (e.g., one or more processors) and a memory 332. The memory 332 stores one or more computer application programs 342 or data 344. The memory 332 can be temporary or persistent storage. The program stored in the memory 332 may include one or more modules (not shown in the diagram), each module including a series of instruction operations on the data processing device. Furthermore, the processor 322 may be configured to communicate with the memory 332 and execute the series of instruction operations stored in the memory 332 on the load forecasting device 301.

[0174] The load forecasting device 301 may also include one or more power supplies 326, one or more wired or wireless network interfaces 350, one or more input / output interfaces 358, and / or one or more operating systems 341.

[0175] The steps in the load forecasting method described above can be implemented by the structure of the load forecasting equipment.

[0176] Corresponding to the above method embodiments, the present invention also provides a computer-readable storage medium storing a computer program, which, when executed by a processor, can perform the following steps:

[0177] Empirical mode decomposition (EMD) is used to decompose the target load power data with mixed noise components to obtain a detrended subsequence of the target quantity. The range of K-value parameters for variational mode decomposition is determined based on the target quantity. The optimal decomposition parameters for variational mode decomposition are determined by combining each detrended subsequence and the range of K-value parameters. The load value corresponding to the target load power data is predicted based on the optimal decomposition parameters for variational mode decomposition.

[0178] The computer-readable storage medium may include various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0179] For a description of the computer-readable storage medium provided by the present invention, please refer to the above method embodiments; the present invention will not be described in detail here.

[0180] The various embodiments in this specification are described in a progressive manner, with each embodiment focusing on its differences from other embodiments. Similar or identical parts between embodiments can be referred to interchangeably. For the apparatuses, devices, and computer-readable storage media disclosed in the embodiments, since they correspond to the methods disclosed in the embodiments, the descriptions are relatively simple; relevant parts can be referred to the method section.

[0181] This article uses specific examples to illustrate the principles and implementation methods of the present invention. The descriptions of the above embodiments are only for the purpose of helping to understand the technical solutions and core ideas of the present invention. It should be noted that for those skilled in the art, several improvements and modifications can be made to the present invention without departing from the principles of the present invention, and these improvements and modifications also fall within the protection scope of the claims of the present invention.

Claims

1. A load forecasting method, characterized in that, include: Empirical mode decomposition is used to decompose the target load power data with mixed noise components to obtain the detrended subsequence of the target quantity; The range of K-value parameters for variational mode decomposition is determined based on the target quantity; The optimal decomposition parameters for variational mode decomposition are determined by combining the detrended subsequences and the range of K values. Predict the load value corresponding to the target load power data based on the optimal decomposition parameters of the variational mode decomposition. The determination of the optimal decomposition parameters for variational mode decomposition, combining the detrended subsequences and the range of K values, includes: Summation is performed on each of the detrended subsequences to obtain the sum sequence; Each positive integer in the range of K values ​​is determined as a target decomposition parameter; For each target decomposition parameter, the sum sequence is decomposed into subsequences corresponding to the target decomposition parameter. Calculate the decomposition residuals based on the sum sequence and each of the subsequences; Calculate the complexity index of the decomposed residual; Determine the maximum value among the various complexity indices, and determine the target decomposition parameter corresponding to the maximum value as the optimal decomposition parameter of the variational mode decomposition. Accordingly, predicting the load value corresponding to the target load power data based on the optimal decomposition parameters of the variational mode decomposition includes: Obtain the target subsequences obtained by decomposing the sum sequence using the optimal decomposition parameters of the variational mode decomposition; Each of the target subsequences is input into a preset load prediction model to perform load prediction and obtain the load value corresponding to the target load power data.

2. The load forecasting method according to claim 1, characterized in that, The range of K-value parameters for variational mode decomposition is determined based on the target quantity, including: The positive integers between two and the target number are defined as the range of K-value parameters for the variational mode decomposition.

3. The load forecasting method according to claim 1, characterized in that, Empirical mode decomposition (EMD) is used to decompose the target load power data of mixed noise components, including: The target load power data of the mixed noise component is sampled at a preset time interval to obtain the power data of each sampled load. The empirical mode decomposition is used to decompose each of the sampled load power data.

4. The load forecasting method according to claim 3, characterized in that, After sampling the target load power data of the mixed noise components at preset time intervals to obtain the sampled load power data, the process also includes: Determine if there is a target time point where the sampling result is empty; If so, retrieve the target sampled load power data corresponding to the target time point in the previous sampling period; The target sampled load power data is determined as the sampled load power data corresponding to the target time point in this sampling period.

5. A load forecasting device, characterized in that, include: The data decomposition module is used to decompose the target load power data with mixed noise components using empirical mode decomposition to obtain the target number of detrended subsequences. The parameter range determination module is used to determine the range of K-value parameters for variational mode decomposition based on the target quantity. The optimal decomposition parameter determination module is used to determine the optimal decomposition parameters of variational mode decomposition by combining each of the detrended subsequences and the range of the K-value parameter. The load forecasting module is used to predict the load value corresponding to the target load power data based on the optimal decomposition parameters of the variational mode decomposition. The optimal decomposition parameter determination module includes: The sequence summation calculation submodule is used to sum the detrended subsequences to obtain the sum sequence; The decomposition parameter determination submodule is used to determine each positive integer in the range of K-value parameters as the target decomposition parameter; The sequence decomposition submodule is used to decompose the sum sequence into subsequences of the target decomposition parameter for each target decomposition parameter; A decomposed residual calculation submodule is used to calculate the decomposed residual based on the sum sequence and each of the subsequences; The complexity index calculation submodule is used to calculate the complexity index of the decomposed residual; The optimal decomposition parameter determination submodule is used to determine the maximum value among the various complexity indicators, and to determine the target decomposition parameter corresponding to the maximum value as the optimal decomposition parameter of the variational mode decomposition. The load forecasting module includes: The subsequence acquisition submodule is used to acquire each target subsequence obtained by decomposing the sum sequence through the optimal decomposition parameters of the variational mode decomposition; The load forecasting submodule is used to input each of the target subsequences into a preset load forecasting model, so as to use the preset load forecasting model to perform load forecasting and obtain the load value corresponding to the target load power data.

6. A load forecasting device, characterized in that, include: Memory, used to store computer programs; A processor for executing the computer program to implement the steps of the load forecasting method as described in any one of claims 1 to 4.

7. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program that, when executed by a processor, implements the steps of the load forecasting method as described in any one of claims 1 to 4.

Images