CN-115374989-B - Load prediction method, device, equipment and computer readable storage medium

CN115374989BCN 115374989 BCN115374989 BCN 115374989BCN-115374989-B

Abstract

The invention discloses a load prediction method which comprises the steps of decomposing target load power data of mixed noise components by adopting empirical mode decomposition to obtain a trend removal subsequence of target quantity, determining a K value parameter range of variation mode decomposition according to the target quantity, determining optimal decomposition parameters of variation mode decomposition by combining the trend removal subsequences and the K value parameter range, and predicting load values corresponding to the target load power data based on the optimal decomposition parameters of the variation mode decomposition. By applying the load prediction method provided by the invention, the effective noise reduction of the load power data of the mixed noise component is realized, and the accuracy of load prediction is greatly improved. The invention also discloses a load prediction device, equipment and a storage medium, which have corresponding technical effects.

Inventors

WANG YUANYUAN
SUN SHANFENG
LUO XIAOMIN
CAI YE
HUANG JINGJIE

Assignees

长沙理工大学

Dates

Publication Date: 20260505
Application Date: 20211213

Claims (7)

1. A load prediction method, comprising: decomposing target load power data of the mixed noise component by adopting empirical mode decomposition to obtain a trending subsequence of target quantity; Determining a K value parameter range of variation modal decomposition according to the target quantity; Determining optimal decomposition parameters of variation modal decomposition by combining each trending subsequence and the K value parameter range; predicting a load value corresponding to the target load power data based on the variation modal decomposition optimal decomposition parameter; Wherein determining a variational modal decomposition optimal decomposition parameter in combination with each of the detrending subsequences and the K-value parameter range comprises: summing up and calculating each trend-removing subsequence to obtain a sum sequence; Respectively determining each positive integer in the K value parameter range as a target decomposition parameter; Decomposing the sum sequence into the target decomposition parameter subsequences for each target decomposition parameter; calculating a decomposition residual according to the sum sequence and each subsequence; calculating a complexity index of the decomposition residual; Determining the maximum value in each complexity index, and determining a target decomposition parameter corresponding to the maximum value as the optimal decomposition parameter of the variation modal decomposition; correspondingly, predicting the load value corresponding to the target load power data based on the variation modal decomposition optimal decomposition parameter comprises the following steps: obtaining each target subsequence of the sum sequence obtained by decomposing the optimal decomposition parameters of the variation mode decomposition; and inputting each target subsequence into a preset load prediction model, so as to utilize the preset load prediction model to perform load prediction, and obtaining a load value corresponding to the target load power data.
2. The load prediction method according to claim 1, wherein determining a K-value parameter range of the variant modal decomposition from the target number comprises: and determining a positive integer between two and the target number as a K value parameter range of the variation modal decomposition.
3. The load prediction method according to claim 1, wherein decomposing the target load power data of the mixed noise component using empirical mode decomposition comprises: sampling the target load power data of the mixed noise component according to a preset time interval to obtain each sampled load power data; And decomposing each sampled load power data by adopting the empirical mode decomposition.
4. The load prediction method according to claim 3, further comprising, after sampling the target load power data of the mixed noise component at a predetermined time interval to obtain each sampled load power data: Judging whether a target time point with a sampling result being empty exists or not; If yes, the target sampling load power data corresponding to the target time point in the previous sampling period is called; And determining the target sampling load power data as sampling load power data corresponding to the target time point in the sampling period.
5. A load predicting apparatus, comprising: The data decomposition module is used for decomposing the target load power data of the mixed noise components by adopting empirical mode decomposition to obtain a trending subsequence of the target number; The parameter range determining module is used for determining a K value parameter range of the variation modal decomposition according to the target quantity; The optimal decomposition parameter determining module is used for determining a variation modal decomposition optimal decomposition parameter by combining each trending subsequence and the K value parameter range; the load prediction module is used for predicting a load value corresponding to the target load power data based on the variation modal decomposition optimal decomposition parameter; Wherein, the optimal decomposition parameter determining module comprises: the sequence summation calculation sub-module is used for carrying out summation calculation on each trending subsequence to obtain a summation sequence; The decomposition parameter determining submodule is used for respectively determining each positive integer in the K value parameter range as a target decomposition parameter; A sequence decomposition sub-module for decomposing the sum sequence into the target decomposition parameter sub-sequences for each target decomposition parameter; a decomposition residual calculation sub-module for calculating a decomposition residual according to the sum sequence and each sub-sequence; A complexity index calculation sub-module, configured to calculate a complexity index of the decomposition residual; the optimal decomposition parameter determination submodule is used for determining the maximum value in each complexity index and determining the target decomposition parameter corresponding to the maximum value as the optimal decomposition parameter of the variation modal decomposition; the load prediction module includes: The subsequence acquisition sub-module is used for acquiring each target subsequence obtained by decomposing the optimal decomposition parameters of the sum sequence through the variation mode decomposition; The load prediction sub-module is used for inputting each target sub-sequence into a preset load prediction model so as to perform load prediction by using the preset load prediction model and obtain a load value corresponding to the target load power data.
6. A load predicting apparatus, characterized by comprising: A memory for storing a computer program; Processor for implementing the steps of the load prediction method according to any one of claims 1 to 4 when executing said computer program.
7. A computer readable storage medium, characterized in that the computer readable storage medium has stored thereon a computer program which, when executed by a processor, implements the steps of the load prediction method according to any of claims 1 to 4.

Description

Load prediction method, device, equipment and computer readable storage medium Technical Field The present invention relates to the field of data noise reduction technologies, and in particular, to a load prediction method, apparatus, device, and computer readable storage medium. Background The specific gravity of the industrial load is the first in the electricity utilization constitution, and for industrial enterprises, a reasonable electricity purchasing plan is formulated according to a load prediction curve of the enterprise for a period of time in the future, so that the waste can be reduced, the electricity utilization cost is reduced, the management on the electricity demand side is realized, and the contradiction between supply and demand of an electric power system is relieved. The accuracy of load power data is greatly influenced by noise, the prediction accuracy of industrial enterprise load prediction is reduced, the noise content of the industrial enterprise load data is reduced, and the method has important significance in improving the industrial enterprise load prediction accuracy. The load power data of the mixed noise components can be decomposed and reconstructed to well filter the noise components contained in the load power data, and the current load power data decomposition method mainly comprises empirical mode decomposition (EMPIRICAL MODE DECOMPOSITION, EMD), variational mode decomposition (Variational Mode Decomposition, VMD) and the like. The non-stationary signal can be decomposed into a component signal with a higher stationary state by using the empirical mode decomposition, but the empirical mode decomposition is easy to generate an endpoint effect and a mode aliasing phenomenon. The variational modal decomposition is a novel decomposition method aiming at non-stationary signals, which improves the defects of the empirical modal decomposition, but the method also has the defect that the value of the decomposition parameter K cannot be determined in a self-adaptive manner. The choice of the K value has a great influence on the decomposition performance of the variation modal decomposition, and in theory, K cannot be too small or too large. For the former, a small number of subsequences K is difficult to fully express the signal, some important parts of the signal can be discarded as noise, and the reconstructed signal after decomposition processing can be severely distorted. For the latter, an excessive number of subsequences K will generally result in excessive decomposition, which not only cannot effectively filter out noise components, but also results in the occurrence of modal aliasing. The load power data of the mixed noise component cannot be effectively reduced in noise, and the accuracy of load prediction is affected. In summary, how to effectively solve the problems that the load power data of the mixed noise component cannot be effectively reduced in noise, and the accuracy of load prediction is affected is an urgent need of those skilled in the art. Disclosure of Invention The invention aims to provide a load prediction method which realizes effective noise reduction on load power data of mixed noise components and greatly improves the accuracy of load prediction, and provides a load prediction device, equipment and a computer readable storage medium. In order to solve the technical problems, the invention provides the following technical scheme: A load prediction method, comprising: decomposing target load power data of the mixed noise component by adopting empirical mode decomposition to obtain a trending subsequence of target quantity; Determining a K value parameter range of variation modal decomposition according to the target quantity; Determining optimal decomposition parameters of variation modal decomposition by combining each trending subsequence and the K value parameter range; and predicting a load value corresponding to the target load power data based on the variation modal decomposition optimal decomposition parameter. In a specific embodiment of the present invention, determining the K-value parameter range of the variant modal decomposition according to the target number includes: and determining a positive integer between two and the target number as a K value parameter range of the variation modal decomposition. In one embodiment of the present invention, determining the optimal decomposition parameters for decomposition of the variation modality by combining each of the detrending subsequences and the K-value parameter range includes: summing up and calculating each trend-removing subsequence to obtain a sum sequence; Respectively determining each positive integer in the K value parameter range as a target decomposition parameter; Decomposing the sum sequence into the target decomposition parameter subsequences for each target decomposition parameter; calculating a decomposition residual according to the sum sequence and each subsequence; calcul